Seismica

Paleoseismic history of the causative faults of the 2019 Ridgecrest, California earthquake sequence

2026-01-16T08:11:44-05:00

The July 2019 Ridgecrest sequence ruptured two nearly orthogonal faults, the left-lateral, NE-striking Salt Wells Valley fault (M_w6.4) and the right-lateral, NW-striking Paxton Ranch fault (M_w7.1), highlighting the hazard of multi-fault earthquakes in the Walker Lane. To test whether similar conjugate ruptures occurred previously, we excavated five paleoseismic trenches and constrained paleo-earthquake timing using luminescence ages. Salt Wells Valley exposures record the 2019 rupture and only one earlier surface-faulting earthquake (17-27 ka), indicating infrequent activity. In contrast, Paxton Ranch strata preserve two Holocene events (4.4-8.7 ka and 10.6-14.6 ka) and up to three late Pleistocene events (17 ka and older) in addition to 2019. These records indicate that rupture along the Paxton Ranch fault commonly occurs independently of the Salt Wells Valley fault. Slip rates based on 2019 displacements and these event intervals are 0.2-1.3 mm yr⁻¹ for Paxton Ranch and 0.01-0.09 mm yr⁻¹ for Salt Wells Valley. The 2019 sequence therefore represents an unusual pairing of an often-active dextral fault with a much less active sinistral fault. The contrasting recurrence and lack of overlap, together with regional paleoseismic patterns, indicate that synchronous rupture is not systematic but instead varies among faults within an evolving network. This suggests that seismic hazard reflects a fault system in which rupture is governed by time-dependent fault network interactions rather than independent, repeatable behavior of individual faults.

Dense Seismic Array Monitoring of the Glacier Tongue of Isunnguata Sermia, West Greenland

2026-07-17T09:48:54-04:00

Accelerating mass loss from the Greenland Ice Sheet is affected by meltwater-driven changes in ice dynamics, which remain poorly understood due to limited observations. Here, we present a 2.5 km² wide dense passive seismic array experiment conducted in the ablation zone of Isunnguata Sermia, West Greenland. We target varying surface melt conditions through one-month long monitoring periods in spring, summer, and fall using 82-117 nodes deployed in 2023 and 2024 complemented by multi-week surface Distributed Acoustic Sensing acquisitions in 2024. We assess data quality using power spectral densities and noise correlation functions. We find that low-frequency seismic power is highly correlated to ice surface velocity, suggesting a strong control of subglacial hydrology on ice dynamics. We retrieve stable and high signal-to-noise ratio noise correlations containing Rayleigh, Love, and P wave arrivals, suggesting these may successfully be used for glacier structure imaging and monitoring. We finally demonstrate we can locate numerous seismic events with resolution down to a few meters using Matched Field Processing and which exhibit characteristic spatial patterns evolving across seasons. These findings establish the potential of such experiment to infer glacier hydrology, dynamics, and structure at high spatial and temporal resolution.

Quantifying the erasure of earthquakes in desert landscapes

2026-07-14T04:58:07-04:00

Seismic hazard analysis depends in part on understanding fault segmentation and slip distribution, which are partially recorded in the landscape during surface-rupturing earthquakes. Over time, surface processes degrade these features, challenging fault mapping. We use landscape evolution models to quantify this information loss in desert environments. Using post-earthquake lidar from the 2019 Ridgecrest (California) and 2010 El Mayor-Cucapah (Baja California) ruptures, we simulate landscape degradation using 2D linear diffusion in Landlab over 100, 1k, 5k, and 10k years, with a transport rate of 1 m²/kyr. We assess change in mappable fault trace length, fault zone width, and a “degradation coefficient” based on topographic slope change. Field validation in 2024 (Ridgecrest) supports the modeled degradation. Results show that 20–80% of original fault trace length remains after 10k years and fault zone width decreases from a mean of 30 m to ~2 m, causing older rupture zones to appear narrower and less complex than initially. Degradation is fastest in the first 100 years, then slows. Fault zones with simple, single-strand morphology retain more mappable length and degrade more slowly. Fault zone structure primarily controls degradation rate and fault trace visibility. These findings provide quantitative constraints on landform degradation, informing probabilistic displacement hazard models and fault mapping in tectonically active regions.

Influence of Moho topography on seismic beamforming: implications for mantle scattering

2026-05-18T11:28:17-04:00

Measurements of scattered seismic phases are used to infer the distribution of small-scale heterogeneities in the mantle, which in turn inform our understanding of mantle dynamics. Directionality of the wavefront can be measured at seismic arrays, and accuracy is essential, as even small deviations can significantly misplace the inferred location of these scatterers. However, near-surface structures under seismic stations may distort these directionality measurements. In this study, we assess how variations in crustal thickness affect array measurements across Alaska. We observe backazimuth offsets mostly within 5° of the great-circle path between source and array, with offsets varying systematically between sub-arrays for each event, indicating a receiver-side origin. We test the hypothesis that offsets are caused by P waves interacting with dipping Moho interfaces, and reproduce the observed offsets. We quantify how even modest tilts in the Moho (~6°) can deflect the wavefront and alter apparent arrival direction by observable amounts. These offsets can shift inferred scatterer locations by hundreds of kilometers, illustrating the need to correct for shallow structure in deep Earth imaging studies.

The Impact of Mapper Experience and Data on the Quality of Geomorphic Fault Mapping

2026-07-14T04:58:08-04:00

Mapping of potentially active faults is critical to characterizing earthquake processes and surface rupture hazard. Geologists map tectonic faults from remote sensing datasets by interpreting tectonic landforms formed during past earthquakes. However, evidence for faulting is often ambiguous, incomplete, and challenging to interpret. Conventional wisdom suggests that mapping skill improves with experience. In contrast, prior work has shown that experienced geologists can still miss faults and disagree on fault location. To test whether, when, and to what extent mapping improves with experience, we designed a repeat mapping experiment involving 23 participants, ranging from undergraduate students to senior professionals. Each participant created regional-scale fault maps using pre-earthquake remote sensing data and a systematic, geomorphology-based approach. While professionals slightly outperformed students, some graduate students produced similar quality maps, and all participants missed identifying tectonic landforms indicative of future ruptures. Using participant mapped fault data, we quantify inter-mapper epistemic uncertainty in fault location to be between 55 and 117 m. Our work gives insight into approaches to improve fault mapping and suggests that experience beyond a graduate student level does not significantly improve mapping results.

Improving seismicity monitoring to prepare for CO2 storage in the Horda platform, Norwegian North Sea

2026-07-15T06:19:06-04:00

The Horda Platform in the Norwegian North Sea will host several awarded CO₂ storage licenses. This requires a robust understanding of background seismicity for site characterization and baseline monitoring. Offshore seismic monitoring is challenging when relying mainly on land stations, leading to poor azimuthal coverage and less precise locations. To improve monitoring in the Horda platform, we deployed three broadband ocean bottom seismometers (OBSs) during October 2021 to September 2022. We used ambient noise cross-correlations to validate OBS timing corrections, which also revealed Scholte waves consistent with the presence of soft sediments. We applied a deep learning-based detection algorithm to continuous OBSs and land data, followed by manual review. The resulting catalog, combined with the Norwegian National Seismic Network, reduced the magnitude of completeness to 0.8. Bayesian hierarchical relocation further refined event locations and increased confidence in detecting small earthquakes. The seismicity analysis revealed previously undetected offshore events, mainly north of the Horda Platform. These events are small and spatially scattered, but provide an improved baseline for seismicity characterization. Our work shows that long-term OBS deployment near the injection site enhances passive seismic monitoring plan for an offshore CO₂ storage site.

Seismo-Acoustic Meteoroid Observation Recording Database (SMORD): A Global Dataset and Deep-Learning Phase Picker for Meteoroid-Generated Air-to-Ground Coupled Seismic Waves

2026-07-03T14:07:45-04:00

Meteoroids impacting Earth's atmosphere generate acoustic waves that can couple into the ground and can be recorded by dense, globally distributed seismic networks. Thus, these records complement optical and radar observations, especially since seismic stations also operate in cloudy weather conditions and during daytime. However, open datasets that link meteoroid events to labeled seismic waveforms are scarce, limiting the development of automated detectors for meteoroid-induced seismo-acoustic signals. We introduce the Seismo-acoustic Meteoroid Observation Recording Database (SMORD), compiled by cross-referencing public meteoroid catalogs (International Meteor Organization fireball reports; NASA CNEOS fireball catalog) with seismic archives. Continuous waveforms are manually labeled for the first clear meteoroid-related onset of air-to-ground coupled seismic waves using a three-level pick-quality scheme. SMORD v1.0 contains 310 meteoroid events and 3,295 labeled arrivals across a global station set. Using SMORD labels, we train a PhaseNet picker in SeisBench with station-level splits and augmentation. On test data, the model achieves 91% precision and 94% recall at a 0.5 decision threshold (area-under-curve value 0.89), with median absolute timing error of 0.02~s (90% within c. ±0.3 s). We demonstrate automated onset detection and trajectory reconstruction for an April 2025 Adriatic fireball, highlighting the values of SMORD for rapid post-event analysis.

Fragility Modeling of Precariously Balanced Rocks: Calibration, Benchmarking, and Sensitivity

2026-07-17T13:41:38-04:00

Precariously balanced rocks (PBRs) provide natural geological indicators for constraining the upper bounds of earthquake ground motions over long timescales. However, translating these constraints into fragility models remains challenging because computationally expensive simulations limit inverse analysis and calibration of contact physics, reducing confidence in the fragility model predictions. To address the challenges, we present a simulated shake-table platform built on a physics engine. Using physical large-scale shake-table experiments on a natural PBR and 582 recorded earthquake displacement histories, we calibrate the contact parameters and benchmark overturning predictions against experimental results and a state-of-the-art discrete element method (DEM). The physics-engine approach reproduces overturning with predictive reliability comparable to DEM, while reducing wall-clock cost by approximately 10² to 10⁵ times. This efficiency enables large ensemble analyses and allows us to evaluate how uncertainty in contact parameters propagates into inferred fragility boundaries. Among the contact parameters, lateral friction exerted the strongest influence on PBR fragility, whereas restitution and spinning friction had comparatively minor effects, with contact damping and stiffness exhibiting more complex behavior. Our study establish a practical pathway for using PBRs as quantitative constraints in seismic hazard assessment.

Integrated seismic monitoring reveals subsurface evolution during volcanic inflation at Askja volcano, Iceland

2026-07-13T07:07:18-04:00

Identifying volcanic transitions from quiescence to unrest and tracking subsurface evolution remains critically challenging. We analyzed continuous seismic records from Askja volcano (Iceland) from 2008-2024 using coda wave interferometry to track relative seismic velocity variations (dv/v) as a proxy for subsurface changes during renewed volcanic intrusion. This analysis is complemented with three additional methods: network covariance matrix analysis, Displacement Seismic Amplitude Ratio (DSAR) single-station analysis, and sensitivity kernel analysis, alongside GNSS, earthquake catalog, and meteorological datasets. Since August 2021, dv/v measurements revealed an abrupt transition from regular seasonal oscillations (±0.2%) to predominantly negative values reaching -0.7% near the inflation center, coinciding with 76 cm GNSS-detected ground uplift by August 2024. DSAR indicated higher seismic attenuation near the inflation source, suggesting volatile accumulation in the shallow subsurface through June 2022, followed by marked decreases associated with depressurization. Sensitivity kernel analysis demonstrated wave sensitivity reaches down to 3 km depth, encompassing the shallow reservoir levels. Integrated seismic observations revealed magma-induced seismic velocity drops, followed by system reorganization with frequency-dependent recovery, and finally establishment of a new state with hydrothermal circulation maintaining altered seismic properties. This approach demonstrates the effectiveness of continuous seismic monitoring for detecting volcanic unrest transitions and tracking evolving subsurface processes.

Crustal Thickness Variations Beneath the Western Indian Ocean Using Teleseismic P-Wave Coda Autocorrelations on Ocean-Bottom Seismic Data

2026-07-03T14:07:53-04:00

The western Indian Ocean is a key region for investigating lithospheric evolution, as it records a complex interplay of tectonic, magmatic, and mantle processes. Constraining crustal thickness across this area is therefore essential for understanding how these processes interact and shape the region's geodynamic development. In this study, we apply teleseismic P-wave coda autocorrelation to map crustal thickness across the western Indian Ocean using data from 54 ocean-bottom seismometers (OBSs) and 7 land-based seismic stations. Our results reveal pronounced lateral variations in crustal thickness, ranging from ~4.3 km beneath young oceanic crust near the Central Indian Ridge (CIR) to ~25.85 km along the eastern margin of Madagascar. The oceanic domain exhibits a mean crustal thickness of ~7.01 ± 0.27 km, consistent with global oceanic averages. Volcanic islands within the Mozambique Channel show crustal thicknesses between ~11.21 and 23.98 km, whereas those in the Mascarene Basin display values of ~10.73 km and ~14.63 km. These localized zones of crustal thickening beneath volcanic islands likely reflect long-lived magmatic underplating and hotspot-related intrusions. Collectively, these findings provide new quantitative constraints on the tectono-magmatic processes that govern crustal formation, modification, and isostatic compensation in this geodynamically complex region.