Allen Husker

The 2017 Mw8.2 Tehuantepec earthquake is the largest instrumentally recorded normal faulting earthquake in Mexico. It occurred offshore on a reactivated outer rise fault within the Tehuantepec megathrust seismic gap that is within the undefined intersection of the North American – Caribbean – Cocos plates. We analyze a relocated machine learning generated aftershock catalog that used data from temporary and permanent seismic stations around the Gulf of Tehuantepec to further interpret the rupture process of this earthquake. Similar to other results, we find that the earthquake initiated 100 km offshore, running northwest 90 km into the intersection. Our findings then show that the rupture jumped 10 km inland as it crossed the trench-perpendicular Tehuantepec Ridge and continued another 60 km before stopping. We used finite fault inversions to distinguish between different possibilities for how the fault ruptured. A two-plane normal fault model offset at ridge, best represents the slip. Moment tensors reveal that the first section before the Ridge was highly complex with a large mix of moment tensors. After the jump across the Ridge, the earthquakes are solely thrust mechanism for 60 km. This indicates that the normal slip on the second half of the fault rupture probably over-slipped and the aftershocks represent the stress accommodation. There was a separate N-S trending divergent region at the location where the Tehuantepec earthquake stopped. We interpret this to be a break between the remaining sliver of the Caribbean plate and the North American plate.

The 8 September 2017 Mw 8.2 Tehuantepec earthquake was the largest instrumentally recorded normal-faulting earthquake in Mexico. The mainshock occurred offshore within the Tehuantepec seismic gap, generating >30,000 aftershocks in the following year. We applied an open-source, machine learning (ML)–assisted workflow to construct a high-resolution aftershock catalog using data from temporary and permanent seismic networks in southern Mexico. The workflow integrates PhaseNet for phase detection; GaMMA for phase association; and VELEST, HypoInverse, and HypoDD for velocity modeling and relocation. We processed seven months of continuous waveform data from 29 broadband stations, including a temporary rapid-response deployment that improved station coverage of the offshore rupture zone. To evaluate performance, we compared our results against analyst-reviewed picks and event locations from the Servicio Sismológico Nacional catalog. The resulting catalog contains 11,374 relocated earthquakes and represents the most comprehensive published dataset for this sequence, incorporating the first full use of the temporary network. Relocated hypocenters show improved depth control and align well with the Slab2.0 subduction geometry, revealing clearer separation between offshore slab events and onshore crustal seismicity. This study demonstrates that combining ML-based detection with established methods provides a scalable and reproducible approach for constructing high-quality earthquake catalogs in tectonically complex environments and offers practical guidance for adapting similar workflows to other earthquake sequences.

We introduce a modular software framework designed to integrate distributed acoustic sensing (DAS) data into operational earthquake monitoring systems. Building on the infrastructure of the Advanced National Seismic System (ANSS) and the Southern California Seismic Network, which employs the ANSS Quake Monitoring Software (AQMS), our solution supports real-time DAS waveform streaming and machine-learning-based travel-time picking to leverage the dense spatial sampling of DAS arrays. To enable seamless compatibility with the AQMS, our approach uses standardized seismic data formats to incorporate predetermined DAS channels. We demonstrate the integration of data from a 100-km-long DAS array deployed in Ridgecrest, California, and provide a detailed description of the software components and deployment strategy. This work represents a step toward incorporating DAS into routine seismic monitoring and opens new possibilities for real-time hazard assessment using fiber-optic networks.

This research expands on work in the field of Distributed Acoustic Sensing (DAS), a new technology that exploits unused optical fiber infrastructure for seismic strain sensing. When deployed on telecommunication cables, particularly offshore, the method has the potential to significantly enhance the coverage of Earthquake Early Warning (EEW) systems. However, while current EEW systems are designed to receive measurements of peak ground acceleration (PGA) from conventional seismic stations, the correlation between DAS amplitudes and ground motion is still not fully understood. There exists a need for more observations to validate this relationship before DAS technology can be included in EEW operations. In this study, DAS peak strain-rate (PSR) values are converted into ground motion predictions by estimating the apparent wave speed using a slant-stack approach. These predictions are compared with real measurements of PGA from co-located seismic stations from the Community Seismic Network to observe correlation and evaluate the accuracy of the method. Results provide new insight into the relationship between DAS amplitudes and ground motion and demonstrate the potential for DAS measurements to augment existing EEW systems.

The use of fiber-optic sensing systems in seismology has exploded in the past decade. Despite an ever-growing library of ground-breaking studies, questions remain about the potential of fiber-optic sensing technologies as tools for advancing if not revolutionizing earthquake-hazards-related research, monitoring, and early warning systems. A working group convened to explore these topics; we comprehensively examined the application of fiber optics in various aspects of earthquake hazards, encompassing earthquake source processes, crustal imaging, data archiving, and technological challenges. There is great potential for fiber-optic systems to advance earthquake monitoring and understanding, but to fully unlock their capabilities requires continued progress in key areas of research and development, including instrument testing and validation, increased dynamic range for applications focused on larger earthquakes, and continued improvement in subsurface and source imaging methods. A key current stumbling block results from the lack of clear data archiving requirements, and we propose an initial strategy that balances data volume requirements with preserving key data for a broad range of future studies. In addition, we demonstrate the potential for fiber-optic sensing to impact monitoring efforts by documenting the data completeness in a number of long-term experiments. Finally, we outline the features of a instrument testing facility that would enable progress toward reliable and standardized distributed acoustic sensing data. Overcoming these current obstacles would facilitate progress in fiber-optic sensing and unlock its potential application to a broad range of earthquake hazard problems.

Machine learning models are a relatively new system for automation of data monitoring and are now being considered for the typically time intensive task of evaluating seismic data in real time. Seismic networks currently possess automated real-time systems that pick phases and detect events as they occur. Newer machine learning models, however, could potentially improve the systems we have now, by increasing the pick accuracy and number of picks, improving event locations, detecting more low magnitude events, and decreasing the workload of analysts. The machine learning model chosen for this study is PhaseNet, a deep-learning neural-net model that was trained on Northern California data, primarily small magnitude earthquakes (3.0 and below), and is currently used in post-processing within the Southern California Seismic Network. Before PhaseNet can be used in real-time monitoring, either in tandem with other systems or by itself, its performance with moderate-to-large magnitude earthquakes must be evaluated. To do this, I have compared phase picks from 50 Southern California earthquakes with magnitudes of 3 to 7.2 that were made by PhaseNet, the current real-time system (AQMS/Earthworm), and seismic analysts. For this analysis, I mainly focused on arrival time differences and whether PhaseNet was able to make a pick using the analyst picks as a ground-truth. Preliminary results indicate that on average PhaseNet makes fewer picks within a short distance of an earthquake, as was expected based on experience from previous applications. In addition, PhaseNet on average seems to pick events slightly earlier in the waveform than analysts for all analyzed earthquakes. I also compare the performance of the current real time system and PhaseNet, as well as looking for any performance trends relating to magnitude and location. Overall, the analysis suggests that PhaseNet could be useful in real-time seismic monitoring when used in conjunction with other pickers or potentially after being retrained specifically for larger magnitude earthquakes.

We refine the newly developed seismicity catalog of the Apollo 17 Lunar Seismic Profiling Experiment, containing over 12,000 thermal events, using quality control algorithms and classify these events by possible source using cross‐correlation, azimuth calculation, and periodicity analysis. We first sort these events into two types—“repeating” or “isolated”—based on whether or not they share similar waveforms with at least one other event. Repeating events are characterized by minimal time between signal onset and maximum, a prevalence during early sunrise, a period consistent with the diurnal cycle, and a preferential direction to the east, coincident with the location of the Lunar Module. Isolated events, on the other hand, are distributed throughout the lunar day, originate from a wide variety of directions, and are characterized by a more gradual increase in amplitude. Many repeating events are likely caused by thermal expansion of the Lunar Module or of enclosed volatiles, while most isolated events may be caused by thermal fracturing of rock or sliding of regolith along crater slopes. In particular, we pinpoint several boulders at the Apollo 17 site whose thermally induced fractures may be the source of 45 moonquakes. Future lunar missions, such as deployments of seismic instruments or gravitational wave detectors, should be aware of thermally induced, repeating seismic signals caused by nearby boulders and structures.

Distributed Acoustic Sensing (DAS) is an emerging technology that converts optical fibers into dense arrays of strainmeters, significantly enhancing our understanding of earthquake physics and Earth's structure. While most past DAS studies have focused primarily on seismic wave phase information, accurate measurements of true ground motion amplitudes are crucial for comprehensive future analyses. However, amplitudes in DAS recordings, especially for pre‐existing telecommunication cables with uncertain fiber‐ground coupling, have not been fully quantified. By calibrating three DAS arrays with co‐located seismometers, we systematically evaluate DAS amplitudes. Our results indicate that the average DAS amplitude of earthquake signals closely matches that of co‐located seismometer data across frequencies from 0.01 to 10 Hz. The noise floor of DAS is comparable to that of strong‐motion stations but higher than that of broadband stations. The saturation amplitude of DAS is adjustable by modifying the pulse repetition rate and gauge length. We also demonstrate how our findings enhance the understanding of fiber‐optic seismology and its implications for natural hazard mitigation and Earth structure imaging and monitoring. Specifically, our results suggest that with proper settings, DAS can detect P‐waves from an M6+ earthquake occurring 10 km from the cable without saturation, indicating its viability for earthquake early warning. Through quantitative comparison and analysis, we also find that local ambient traffic noise levels strongly affect the quality of seismic interferometry measurement, which is a powerful tool for near‐surface imaging and monitoring. Our methodology and findings are valuable for future DAS experiments that require precise seismic amplitude measurements.

The ShakeAlert earthquake early warning (EEW) system partners along with U.S. Geological Survey (USGS) licensed operators deliver EEW alerts to the public and trigger automated systems when a significant earthquake is expected to impact California, Oregon, or Washington. ShakeAlert’s primary goal is to provide usable warning times before the arrival of damaging shaking. EEW is most likely to achieve this goal in large-magnitude earthquakes. In recent years, ShakeAlert has gone through a series of upgrades to its underlying scientific algorithms aimed at improved performance during large earthquakes. Version 3 of this software recently went live in the production system and includes improvements to all algorithms. The main seismic algorithms that detect an earthquake and characterize its location, magnitude, and fault rupture orientation are faster than older versions. Other key changes include: using real-time geodetic data to characterize the magnitude growth in large earthquakes; the introduction of an alert pause procedure to compromise between speed near the epicenter and improved accuracy at larger distances; and the inclusion of a nonergodic site-response model in the ground-motion predictions. ShakeAlert has achieved its primary goal of usable warning times before strong shaking at some locations in real-time operations in recent M 6 earthquakes. Using offline tests, we demonstrate usable warning times are possible for many sites with peak shaking values of modified Mercalli intensity (MMI) 7–8 in M 7+ earthquakes and also for many MMI 8–9 sites in M 8+ earthquakes. ShakeAlert partners use a variety of MMI and magnitude thresholds in deciding when to alert their users within bounds set by the USGS. Our study shows that there is room to raise the magnitude thresholds up to about M 5.5 without adversely affecting performance in large earthquakes. The ground-motion criteria are more complex owing to a significant drop-off in warning times between the MMI 4 and 5 levels of predicted shaking. However, widely used ShakeAlert products, such as the MMI 3 and 4 contour products, can provide sufficiently long warning times before strong shaking in moderate-to-great earthquakes to enable a range of protective actions.

We refine the newly developed seismicity catalog of the Apollo 17 Lunar Seismic Profiling Experiment using quality control algorithms, and speculate on the source mechanisms of thousands of catalogued events using cross-correlation and calculation of event and waveform properties. We first sort these events into two types – “repeating’ or “isolated’ – based on whether or not the event shares similar waveforms with at least one other event. Repeating events are characterized by minimal time between signal onset and maximum, a prevalence during early sunrise, a periodicity consistent with the diurnal cycle, and a preferential direction due east, coincident with the Lunar Module. Isolated events, on the other hand, are distributed throughout the lunar day, originate from a wide variety of directions, and are characterized by a more gradual increase in amplitude. Many repeating events are likely caused by thermal expansion of the Lunar Module or its trapped volatiles, while most isolated events may be caused by thermal fracturing of rock or sliding of regolith along crater slopes. Across both the repeating and isolated events, we pinpoint 43 moonquakes whose time-of-day and waveform characteristics are consistent with thermally-induced fracture of several boulders at the Apollo 17 site. Future lunar missions, particularly those sensitive to ground motion, should be aware of thermally-induced, repeating seismic signals caused by nearby boulders and structures.