Geophysical Journal International

SummaryThe intrinsic attenuation of a seismic wave is a key property of rocks. Knowledge of Q (quality factor), which measures the attenuation, provides information not only about the type of rock, but above all about the pore fluids. In addition, the attenuation is needed for inverse Q filtering and is useful for localizing seismic sources. The methods for determining Q are based on the amplitude and frequency content of the signal, e.g. the spectral ratio and frequency shift methods. In this paper, we consider a nearly frequency-independent Q factor so that the attenuation factor is proportional to frequency, so that high frequencies are attenuated. We use the frequency shifts (from the source to the receiver) to estimate the Q-factor for media with arbitrary geometric interfaces. The amplitude can be affected by factors other than attenuation, namely, geometric spreading and transmission coefficient. The shifts depend on the type of spectrum, which differs for seismic, microseismic and seismological (earthquake) sources. The inversion for Q assumes that the source spectrum is known, as well as the seismic velocities and the location of the interfaces and the source. On the other hand, if Q is known, sources can be located based on a generalization of the Battaglia-Aki method to heterogeneous media and using the centroid of the spectrum and the traveltimes at the receivers instead of the amplitude of the signal in the time domain. Alternatively, we provide formulas for the maximum of the signal spectrum that can also be used (peak frequency shift). Modeling and inversion is performed with 2D and 3D ray tracing algorithms based on Fibonacci search and minimizations with the Praxis algorithm and simulated annealing. We generate synthetic test seismograms with a direct 2D full-wave algorithm based on the pseudo-spectral Fourier method, compare the results with those of ray tracing, and use the minimization algorithm in combination with ray tracing to determine the source location.

SummaryMoment magnitude (Mw) is a widely accepted magnitude scale as a direct physical measure of the long-period seismic energy released at the foci, and thus its reliable quantification is of great importance for accurate probabilistic seismic hazard assessment (PSHA) studies. Yet, a robust estimation of Mw and radiated energy (ER) over a wide range of magnitudes is difficult, mainly due to the existing strong lateral heterogeneous nature of the crust. Furthermore, converting short-period magnitudes such as local magnitude (ML) to Mw can often lead to significant bias. To address this issue, we employ a coda envelope-based source spectral method, which depends on a regional empirical calibration approach that can lower the threshold for reliable Mw and ER estimates. To achieve this aim, we analysed horizontal component waveforms recorded at broadband stations operated by the Kandilli Observatory and Earthquake Research Institute and the Disaster and Emergency Management Presidency from 51 selected moderate local and regional earthquakes (ML ≥ 4.0) that occurred between 2013 and 2024 in and around the central North Anatolian Fault Zone (CNAFZ), including the 23 November 2022 Mw 6.0 Düzce and the 18 April 2024 Mw 5.6 Tokat events. The Java-based Coda Calibration Tool (CCT) implemented on these waveforms enabled a successful establishment of the coda-derived source spectrum that allowed us to obtain robust estimates of apparent stress (σA) and Mw across the CNAFZ. Following the calibration with reference events, we extend reliable magnitude estimation to smaller earthquakes (3.5 ≤ ML < 4.0), confirming robust predictions. Beyond providing a more thorough event catalogue in the CNAFZ, our results reveal low σA levels, comparable to those reported in recent studies in the Marmara Sea, and indicate non-self-similar rupture behaviour that may improve future regional seismic hazard assessments. This approach may also serve as a framework for reliable small-to-moderate earthquake analysis in other tectonically active regions in Türkiye, thereby supporting broader seismic risk management efforts.

SummaryThe Thermal Lattice Boltzmann Method (TLBM) offers an alternative to classical PDE based numerical methods for modelling planetary dynamics. It solves the Lattice Boltzmann Equation on a unitary discrete lattice which requires no matrix operations, and hence scales extremely well on HPC clusters. We describe the TLBM equations which are capable of modelling solid state mantle convection and turbulent magma ocean dynamics, and we present how to obtain the physical (dimensional) time scale from the lattice time scale (number of time steps), convert quantities in physical (dimensional) units (η, κ, σ…) to unitary (dimensionless) lattice units, and strong scaling performance in 3D to 300K cores. We present examples to illustrate use of the TLBM to model plate tectonics, 2D and 3D mantle convection, turbulent magma ocean dynamics, and planetary accretion. We present a table of run times which demonstrate the TLBM’s high throughput performance. The TLBM’s performance enables 3D global convection modelling of the whole mantle on a lattice with up to ∼2 × 1011 grid points corresponding to a resolution of 3 km, and magma ocean dynamics modelling into the ultimate turbulent regime of Ra ≳ 1014, to at least Ra = 1015 (demonstrated to date in 2D for an aspect ratio of 1), and projected to be possible up to Ra = 1018 in 2D or Ra = 1015 in 3D for low aspect ratio models. The TLBM is fast and makes it possible to run huge models which opens exascale computing to planetary dynamics research. We believe that the TLBM provides a valuable new means to advance geodynamical research into the future by enabling fast and high-resolution planetary dynamics simulations including of global mantle convection, magma ocean dynamics into the ultimate turbulent regime, and planetary accretion.

SummaryThe Red Sea is one of the youngest ultraslow-spreading ridges on our planet and an ideal place to investigate the transition from continental rifting to oceanic spreading. Within this context, the Zabargad Fracture Zone (ZFZ) stands out as a particularly intriguing region. The ZFZ hosts notable features, including an offset in the Red Sea spreading axis, the Mabahiss Mons submarine volcano, and the Kebrit Deep brine pool. Additionally, this region is seismically active, posing a hazard to nearby coastal communities. Despite previous geophysical studies, few seismic velocity models image shallow subsurface structures in the ZFZ. In this work, we use approximately one year of seismic ambient noise recorded by five broadband stations and 14 ocean-bottom seismometers to estimate the shear-wave velocity structure of the ZFZ. For this, we compute vertical-vertical, radial-radial, and transverse-transverse noise correlations, obtain group- and phase-velocity dispersion curves of Rayleigh and Love waves in the 3 to 12 s period band, and employ transdimensional tomography to estimate 1-D and 3-D isotropic shear-wave velocity models of the ZFZ. Our 1-D velocity model suggests that, on average, the ZFZ crustal structure comprises a 1.5 km thick layer including hemipelagic sediments and evaporites, a 2.8 km thick oceanic basement, and a crust-mantle transition extending from 6.5 to 8 km below sea level. Meanwhile, our 3-D model agrees with previous geological and geophysical observations and reveals new subsurface structures. In particular, it shows low velocity areas to the east and south of Mabahiss Deep that correlate with known sedimentary basins. Moreover, our 3-D model contains a low-velocity area near Kebrit Deep that correlates with a region of low seismic activity and a recently inferred spreading-axis segment. Based on previous evidence of inactive hydrothermal vents, we infer that this low-velocity area indicates higher basement temperatures near Kebrit Deep compared to other areas. Lastly, our 3-D model displays a velocity contrast in the southern ZFZ that correlates with a contrast in free-air gravity anomalies and a gradient in evaporite topography. Based on these observations, we interpret this velocity contrast as a lineament related to folded evaporites. Our findings present new constraints on the crustal structure of the ZFZ and serve as a reference for other young ultraslow-spreading ridges.

SummaryWe present the results of a three-dimensional seismic tomography study of the upper crust beneath a quadrant of the peak ring structure of the Chicxulub meteorite impact crater, Mexico. Reflection and refraction travel-times from a grid of seismic profiles recorded by a 6 km streamer and 48 ocean bottom seismometer stations were inverted to give a well-resolved three-dimensional velocity model to a maximum depth of 6-8 km. The model comprised the thin water layer, a layer of low seismic velocity post-impact sedimentary infill, and the crater basement, which was separated from the fill by the interface representing the top of the crater, defined by normal incidence reflection picks. The crater basement shows a cylinder-shaped feature extending vertically downwards beneath the topographic peak ring to at least 8 km, the depth of resolution of this survey, characterised by slower seismic velocities than in the surrounding rocks at the same depth. This result supports and extends the observations of previous seismic refraction work in the peak ring, and also of scientific drilling, that the material in the peak ring has significantly reduced seismic velocity compared to typical granitic basement lithologies. We used the velocity model to perform a pre-stack depth migration of a key seismic reflection profile. In the best-fitting model presented here the prominent dipping reflector previously identified on seismic reflection profiles, which projects to the outer edge of the peak ring, dips inwards and crosses the low velocity cylinder without an apparent first order contrast in impedance. This result implies that the reflectivity of this dipping reflector is due to a thin, high-contrast layer such as entrapped impact melt or hydrothermal alteration within the overturned structures. We also identify well-imaged slump blocks from the crater rim/inner ring inwards, variations in the height and width of the peak ring, and associated variations in the velocity contrast that characterises the anomaly beneath the peak ring.

SummaryGeodetic measurements of ground deformation are crucial to identifying and interpreting geophysical processes. We develop a method to fuse data streams from multiple geodetic techniques into a single, three-component deformation field with quantified uncertainties, and without invoking a geophysical model. The fusion is formulated as a semi-parametric latent factor model: a linear mapping ties each observation to the underlying 3-D displacement, while the displacement components are represented nonparametrically with multi-output Gaussian process priors. To achieve practical performance at regional scale, we deploy two complementary sparse GP engines: an Informative Vector Machine (IVM) that selects a small, most-informative active set for fast subset-of-data inference, and a Sparse Variational GP (SVGP) that summarizes the full dataset with inducing points and optimizes a global variational bound. Together, these reduce the scaling of the computation to near-linear in data size and cubic only in the active/inducing set, enabling the potential of fusion of dense geodetic data while maintaining rigorous uncertainty quantification. We demonstrate the approach on coseismic deformation from the 2020 Sparta, NC, USA and 2016 Meinong, Taiwan earthquakes, fusing interferometric synthetic aperture radar (InSAR) with either light detection and ranging (LiDAR) or global navigation satellite system (GNSS) data, respectively. The fused solutions show marked improvements in the precision and coherence of the resolved deformation field and deliver robust, spatially explicit uncertainty estimates. The methodology is readily extensible to time-varying observations to produce four-dimensional (space-time) deformation fields, offering a scalable path to richer characterization of transient geophysical phenomena.

SummaryThe 2019 Mw 6.4 Durrës earthquake in Albania caused severe loss of life and economic damage, highlighting the seismic hazard along the Adriatic–European plate boundary. This study provides the first high-resolution analysis of the aftershock sequence based on data from a dense local seismic network deployed three weeks after the mainshock. Using machine-learning detection and phase-picking tools, we identified 19 152 aftershocks (Ml − 1.8 to 4.6; Mc ≈ 1) over a nine-month period. Based on a newly derived 1D velocity model with station corrections, accounting for large vertical and lateral velocity variations, we relocated the events applying cross-correlation based differential travel times and the double-difference algorithm. The refined seismicity images clearly reveal several sub-parallel ∼30° NE-dipping blind fault structures; the most prominent one, between 12 and 18 km depth, probably hosted the Durrës mainshock. The blind thrust faults lie beneath thick sediments and cut through a carbonate platform and into the Adriatic basement, indicating thick-skinned deformation. Our observations may be interpreted as incipient large-scale slicing and underplating of subducted Adriatic crust. Additional shallow seismicity within a duplex structure in the hanging wall points is relevant for seismic hazard, as even a relatively moderate earthquake occurring close to the surface could cause significant damage.

SummarySeismograms contain a variety of signals in addition to direct waves. The local later phase in the S coda (LPS; strongly reflected or scattered S-waves by subsurface heterogeneities when the hypocentre, station, and the origin of the later phase are located relatively close to each other) is useful for investigating fine-scale subsurface heterogeneities such as significant velocity contrasts and localized scatterers. With the development of dense seismic observation networks, accurate and rapid automatic processing techniques for large numbers of seismograms have become important. The introduction of deep learning techniques enables us to automate the processing of seismograms, such as phase detection and arrival time picking of direct waves with high quality and speed. To utilise the information on LPSs contained in large volumes of seismic waveform data, we developed the LPS-detector, a convolutional neural network-based automatic LPS detection model. We trained the LPS-detector using single-station data from the Moriyoshi volcanic area in northeastern Japan, which is known for observing distinct LPSs. We create an original training dataset with 5 000 earthquakes in this area. The training dataset was separated into the training (3 200), validation (800), and test (1 000) subsets. The LPS-detector yielded a 0.910 area under the curve (AUC) of the receiver operating characteristic (ROC) curve and 0.966 AUC of the precision-recall (PR) curve. These scores indicate good performance for automatic LPS detection, which is comparable to that of manual detection. Additionally, we confirmed that the LPS-detector could detect LPSs at other stations if the waveform characteristics were similar to those of the training subset. The large volume of the LPS catalogue obtained by the LPS-detector provides a new insight into the LPS origin in the Moriyoshi volcanic area. The LPS-detector detected 3 951 new LPSs outside of the training period. Combined with manual detection, 7 599 LPSs were detected in this area. Based on the comprehensive LPS catalogue, we found that the origin of LPSs in this area lies just beneath the earthquake swarm and is inclined at approximately 12° with a south-westward dip. These results indicate that the LPS-detector enables us to extract LPS information from seismogram big data and explore fine-scale subsurface structures.

SummaryThe British Palaeogene Igneous Province records the palaeomagnetic field at a time when it had returned to a reversing state following the Cretaceous Normal Superchron. Whilst the province is well studied for palaeomagnetic directions, palaeointensity results remain scarce. New palaeointensity and palaeomagnetic directional results are presented from a Palaeocene basaltic dyke swarm on the Sleat Peninsula, Skye (NW Scotland). The mean palaeomagnetic direction from 24 dykes has declination 356.7° and inclination 60.0°. The corresponding palaeopole lies at 75.2°N, 181.8°E with associated 95 per cent confidence interval, A95, of 6.6° which is close to, but distinguishable from, previous results from this swarm. The angular dispersion of virtual geomagnetic poles (VGPs), SB, was 17.9 ± 3.2°, consistent with stationary VGP dispersion throughout the Palaeogene. Three experimental palaeointensity methods were attempted: Shaw Double Heating Technique, thermal Thellier, and microwave Thellier. Palaeointensity experiments were successful from two dykes giving results of 15.8 ± 2.3 μT and 12.0 ± 3.4 μT which correspond to virtual dipole moments of 27 ± 4 ZAm2 and 21 ± 6 ZAm2. These are notably weaker than the long-term average and the recent field. Additionally, we present proof-of-concept of a novel method for estimating palaeointensity. The preliminary method was applied to results from the thermal Thellier experiments and is based on reproducing non-ideal Arai plot behaviour using a phenomenological model of thermal remanent magnetisation. This inversion highlighted laboratory field orientation as a major control on Arai plot shape with antiparallel fields predicted and observed to cause major distortions. It produced similar palaeointensity results to the conventional approach but derived from significantly more specimens.

Summary3D magnetotelluric (MT) inversion is a powerful tool for imaging the Earth’s electrical structure, yet its computational demands remain a major challenge, particularly when simulating responses across broad frequency bands. Conventional inversion schemes rely on a unified mesh for forward simulation at all frequencies, which inflates the number of model parameters and greatly extends runtime. To overcome these limitations, we present a finite-volume-based frequency-domain survey decomposition (FVFSD) method that adaptively constructs forward simulation meshes according to the skin depth of each frequency while maintaining a fixed horizontal discretization. This design decouples forward simulation meshes from the inversion mesh, striking a balance between accuracy and efficiency. To further improve the treatment of resistivity contrasts, an equivalent circuit scheme is employed to compute effective conductivities within control volumes, outperforming conventional volume-weighted averaging. We validate the proposed method through comprehensive numerical experiments, including synthetic benchmarks and real-world case studies from the Akebasitao region (China) and the Southern African MT experiment. Results demonstrate that FVFSD achieves accuracy comparable to standard finite-difference forward simulations, while significantly reducing computational time. In large-scale MT inversions, this acceleration directly translates into faster convergence and more efficient recovery of lithospheric-scale resistivity structures. The method is fully compatible with existing inversion algorithms and solvers, making it straightforward to integrate into standard workflows. Overall, FVFSD provides a scalable and accurate strategy for advancing 3D MT inversion, with clear implications for lithospheric studies, resource exploration, and tectonic investigations.

SummaryDistributed Acoustic Sensing (DAS) captures seismic wavefields through precise measurement of phase changes in back-scattered laser pulse signals in an optical fiber that reflect distributed strains or strain rates over continuous fiber segments. The limitation of the DAS wavefield measurement for near surface imaging remains unclear and is worthy of verification with conventional dense nodal arrays. This study compares co-located DAS and dense nodal array recordings in near-surface active surveys conducted in Shenzhen, China. Our field experiment reveals the spatially fixed, step-like distortions of coherent signals in DAS wavefield recordings. These artifacts are attributed to the gauge-length spatial average acting upon fiber segments with heterogeneous sensitivity, which can arise from variations in backscatter intensity and cable coupling conditions. These distortions cause partial deviations between waveforms of DAS and co-located seismometer from their theoretical relationship. A statistical comparison of phase differences between DAS and seismometer waveforms across all nodal points reveals a normal distribution, indicating the prevalence of such distortions in DAS recordings. Moreover, the distortion becomes more severe as frequency increases, resulting in overly dispersed phase difference distributions between DAS and seismometers. This leads to discrepancies in the dispersion curves extracted from them at high frequencies, as dispersion extraction depends on the relative energy peaks contributed by the dominant channels. This study demonstrates that within the frequency range not severely affected by distortion, DAS can effectively extract comparable dispersion curves and velocity structure profiles in place of the nodal array for near-surface imaging applications, as validated through full-profile comparisons. But the DAS wavefield distortions limit advanced applications of the complete wavefield for better subsurface characterization, and imply that DAS data simulation and instrument response analysis should consider channel ensembles.

SummaryA temporary array of seven ocean bottom seismometers (OBS) was deployed offshore the Guerrero subduction zone in Mexico to monitor previously unreported shallow seismicity. These OBS instruments are especially valuable for studying earthquake activity in the Guerrero seismic gap, where a future large event could severely impact densely populated regions of Mexico. This study investigates the shallow seafloor structure, including site effects, shear wave attenuation, and velocity models, using both earthquake data and ambient seismic noise. We employed spectral inversion to estimate the quality factors of shear wave attenuation and site effects. Additionally, we calculated the microtremor horizontal-to-vertical spectral ratio (HVSR) as a proxy for site response and invert it using constraints from hydroacoustic seafloor profiles, parametric sub-bottom profile system (TOPAS), to derive the shallow velocity structure beneath the stations. The inclusion of TOPAS data in the inversion significantly improved convergence, reduced misfit, and resulted in more reliable subsurface models. The HVSR inversions indicate the presence of water-saturated sediments within the upper 250 m, characterized by shear-wave velocities ranging from 55.2 to 1950 m/s and Vp/Vs ratios between 1.80 and 27.84. Strong attenuation effects, typical of marine environments, were observed, with Q(f) values as low as Q = 86f0.62 in the forearc accretionary wedge. Our attenuation estimates are consistent with those found in other offshore subduction zones, contributing to a broader understanding of shallow structures in similar tectonic settings worldwide. We found strong agreement between the estimated site effects and HVSR results, underscoring their close relationship and supporting the reliability of our site response estimates. This is the first study in Mexico to use OBS data to characterize offshore attenuation, site effects, and velocity structure, information that will support future seismological analyses, including earth structure imaging and investigations of both large earthquakes and shallow slow earthquakes in the Guerrero seismic gap.

SummarySeismic sources are typically characterized as stochastic slip distributions on complex fault geometries, which pose significant challenges for computational modelling. Source scaling laws, however, offer a streamlined alternative by correlating simplified fault geometry and slip characteristics with earthquake magnitude. So far, distinct scaling laws have been developed for different tectonic settings and fault mechanisms. However, regional variations in source parameters have not been explicitly quantified. For example, it remains unclear whether earthquakes occurring in similar tectonic environments (e.g. subduction zones) and fault mechanisms (e.g. reverse faulting), but in different regions such as Japan, South America, or Indonesia, exhibit comparable source characteristics, or how such variability should be incorporated into scaling relations. To address this gap, the present study performs a comprehensive exploratory analysis of earthquake source attributes derived exclusively from finite-fault models, including stress drop, alongside standard fault geometry and slip parameters. The analysis spans multiple groupings defined by tectonic setting, fault mechanism, seismic region, focal depth, crustal type, as well as fault-plane inversion modality and spatial resolution, which are examined to account for modelling-related variability across datasets. Stress-drop proxy and slip-parameter estimates, particularly for large magnitude earthquakes, display systematic deviations from self-similar scaling assumptions. Fault-plane modality, defined by the type of seismic and/or geodetic data used in the inversion, and fault-plane resolution, quantified by subfault discretization, are found to be associated with systematic differences in inferred slip and asperity parameters, and help explain part of the intra-event variability observed when multiple models exist for the same earthquake. These factors are therefore incorporated explicitly to isolate physical variability from modelling effects. Based on these findings, existing source scaling laws are revised using a mixed-effects regression framework. Tectonic setting, inversion modality, and fault-plane resolution are treated as fixed effects, while fault mechanism and seismic region are modelled as random and nested-random effects, respectively. The refined scaling relations provide more robust estimates of fault geometry (length, width, area, and asperity dimensions) and slip statistics (mean slip, maximum slip, and slip standard deviation), and are particularly valuable for region-specific computational source modelling and physics-based seismic hazard analysis.

SummaryThe reaction-infiltration instability has been identified as an important mechanism responsible for the formation of high-porosity melt channels in the upper mantle. To better understand this mechanism, we perform linear stability analysis and non-linear numerical simulations in a compacting, chemically reactive porous medium. Strong interactions between compaction and reaction lead to two distinct unstable features: (1) high-porosity channels and (2) compaction-dissolution waves. Here we present a regime diagram to show that, compared to high-porosity channels, compaction-dissolution waves are favoured in systems with lower reaction rate, lower compaction viscosity (i.e., more easily compactible medium), and smaller solubility gradients. This regime diagram predicted by linear stability analysis shows good agreement with the non-linear numerical simulations. Under geologically relevant conditions, both high-porosity channels and compaction-dissolution waves may form in the mantle, although channels are more commonly expected.

SummaryThe Sanriku-Oki subduction region in northeastern Japan is a tectonically active zone where the Pacific Plate subducts beneath the Okhotsk Plate, generating frequent earthquakes. In this study, we present a three-dimensional unstructured S-wave traveltime tomography model of the off-Sanriku forearc using local earthquake data recorded by ocean-bottom Distributed Acoustic Sensing (DAS). The dense spatial sampling provided by DAS enables cost-effective, high-resolution imaging of the forearc region that is difficult to achieve with conventional ocean-bottom seismometer deployments. Eight local earthquakes recorded near the DAS cable provided 209,193 high-quality S-wave arrival times after applying data-selection criteria. These earthquakes were recorded using two DAS interrogator units: the AP Sensing N5200A (70 km coverage, 5 m channel spacing) and the OptaSense QuantX (100 km coverage, 2 m channel spacing). The initial reference models were constructed on a 3D tetrahedral mesh with target cell sizes of 1.5 and 3.0 km, accurately incorporating the DAS cable geometry and earthquake hypocentres. Synthetic traveltimes were efficiently computed using the Fast Sweeping Method (FSM) and the traveltime models were subsequently refined through traveltime inversion. The resulting S-wave tomography models reveal significant spatial heterogeneity within the off-Sanriku forearc region. Low S-wave velocities (Vs ∼ 0.78–0.85km s−1) indicate shallow, unconsolidated Neogene sediments, underlain by higher velocities in the Cretaceous basement (Vs ∼ 1.2–2.4km s−1). The lower crust and uppermost mantle wedge exhibit clear along-strike segmentation of the forearc. In the landward domain, high S-wave velocities in the lower crust (4.15–4.4km s−1) and uppermost mantle (4.6–4.65km s−1) indicate a mechanically strong overriding plate and a thick, cold mantle wedge, with no evidence of significant partial melting or serpentinisation associated with subduction processes. In contrast, the central offshore region exhibits moderately reduced S-wave velocities (3.9–4.1km s−1), suggesting a mechanically weaker overriding plate. Toward the trench, S-wave velocities decrease markedly within the forearc crust, defining a low velocity zone associated with crust–crust interaction above the subducting Pacific Plate and likely reflecting fluid-rich, deformed forearc material in the shallow subduction environment.

SummarySeismic wave speed monitoring is important for the non-destructive evaluation of material properties in response to external forcing. Coda wave interferometry (CWI) uses travel time perturbations in multiply-scattered seismic wave trains – the seismic coda – to detect subtle perturbations in bulk wave speed. However, conventional body-wave CWI cannot separate the coupled contributions of P and S waves, which are sensitive to different material properties. We introduce energy partitioning inversion which decouples these modes by combining a scattering model with CWI measurements within non-equipartitioned coda windows. We applied this methodology to repeated ultrasonic pulse surveys during two laboratory loading experiments on Clashach sandstone: a dynamic experiment (constant strain rate until brittle failure) and a quasi-static experiment (modulating stress to maintain constant acoustic emission rate and slow down the failure process). Relative travel time perturbations and their full covariance between all pairs of surveys were measured across multiple coda windows and inverted for a single perturbation profile using a least-squares method to minimise the variance of the profile. Using an isotropic point scatterer model to predict mode partitioning with respect to the coda lapse time, we invert travel time perturbations for the scattering mean free path travel time and relative P and S wave speed perturbations via Markov-chain Monte Carlo inversion to quantify uncertainty. P and S wave speed perturbations were resolved with 95 % credible intervals of 0.025 % and 0.008 %, respectively. During the quasi-static experiment the temporal resolution was sufficient to capture a quasi-linear decrease in P and S wave speeds by ~ 50 % and ~ 14%, respectively, from peak to failure. The peak P and S wave speed perturbations were ~ 33% lower and ~ 75% higher, respectively, compared to those found in the dynamic experiment. These results demonstrate that CWI and energy partitioning inversion enables the robust, uncertainty-quantified evaluation of separate relative bulk P and S wave speed perturbations in strongly-scattering media.

SummaryThe rigidity and intraplate deformation of the Indian plate has long been a subject of debate. To understand the present-day intraplate deformation, we utilized data from 34 well-distributed continuous Global Positioning System (cGPS) stations across the stable part of the Indian subcontinent to derive refined Euler pole parameters (51.834° ± 0.0495° N,3.9708° ± 0.9351° E, Ω = 0.52067° ± 0.002684°/Myr) and establish an updated Indian Reference Frame (IRF). The intraplate velocity field suggests a westward motion of the Northern Tectonic Block (NTB) (up to 2.5 mm/yr) relative to the stable Southern Tectonic Block (STB) (≤1 mm/yr) along the Narmada-Son Lineament (NSL) in the Central Indian Tectonic Zone (CITZ). The rotation pattern inferred from strain analysis also exhibits a progressive increase in the anticlockwise rotation from the STB (~2 nrad/yr) to NTB (~6 nrad/yr). Thus, the subtle, yet systematic variation in the intraplate deformation pattern from STB to NTB across the NSL strengthen the non-rigid behaviour of the Indian sub-continent and reflects its accommodation of far-field plate boundary forces generated by the Indo-Eurasian collision in the Himalayan region. To further substantiate these findings, we propose to conduct advanced investigations using a denser network of cGPS stations over the Indian subcontinent.

SummaryUncertainty quantification is indispensable for reliable magnetotelluric (MT) data interpretation, given the inherent non-uniqueness of MT inverse problem solutions. However, traditional sampling-based probabilistic schemes often require millions of costly forward predictions, making them computationally prohibitive. To address this, we develop a trans-dimensional Bayesian inversion framework that incorporates a novel forward modelling operator and a reversible-jump Markov chain Monte Carlo (RJMCMC) algorithm for robust uncertainty quantification. The forward modeling method leverages the extended Fourier DeepONet (EFDO) network. Once trained, the EFDO achieves up to a 300-fold acceleration in forward predictions compared to a conventional Finite Volume (FV)-based solver. Furthermore, we utilize an adaptive Delaunay model parameterization during the sampling process to allow for efficient model space exploration. We demonstrate the efficacy of the proposed approach through numerical experiments and application to the COPROD2 MT field dataset. Overall, this work advances Bayesian MT inversion by enabling rapid, high-dimensional inference of subsurface electrical resistivity structures, thereby facilitating reliable geological interpretation.

SummaryDistributed Acoustic Sensing (DAS) has emerged as a valuable complement to conventional seismic monitoring techniques. By converting fiber-optic cables into dense arrays of virtual sensors, DAS enables the application of standard large-array processing methods. However, its directional sensitivity—limited to strain measurements along the fiber axis—may restrict its potential for full wavefield analysis. To address this limitation, we investigate the capabilities of DAS on a fiber-optic cable installed both horizontally, near the surface, and vertically, in a borehole, thereby creating a so-called 3D-DAS array. The survey was carried out in the southern Munich region (Germany) to monitor local seismicity associated with nearby deep geothermal operations. In this study, we present the data acquisition setup and describe a processing workflow developed to characterize source and wavefield parameters of seismic events from DAS recordings. The workflow is illustrated using a nearby ${{M}_w}$ = 0.48 seismic event. Taking advantage of the configuration of the fiber optic cable, we demonstrate that the 3D-DAS array enables estimation of the wavefield back-azimuth, incidence angle and slowness, and compare these results with those provided by a local network of seismometers. In addition, seismic source parameters, including seismic moment and stress drop, are estimated from DAS data acquired in the 250-meter-deep vertical well. These parameters are derived after converting strain-rate to ground motion, a process quantitatively validated using a co-located three-component broadband seismometer. The results and waveform evaluation demonstrate that the 3D-DAS array provides reliable and comprehensive measurements, independently of the existing local seismic network.

SummaryIn many fields of geoscience, researchers study the Earth’s properties by solving inverse or inference problems. Probabilistic approaches have gained increased attention over the past decade because they address the non-linearity and non-uniqueness properties of many naturally-inspired inverse problems and allow uncertainties in the solutions to be estimated. However, implementing such methods is computationally expensive and requires expertise in inverse and inference theory, high performance computing, and the geoscientific theory to be inverted. This makes the methods inaccessible to many geoscientists. In this paper, we first review the theoretical background of a particular suite of probabilistic algorithms referred to as parametric variational inference (PVI), and introduce GeoPVI, an open-source Python package designed to facilitate the implementation of these methods. With GeoPVI, users can model uncertainties in their geophysical parameter estimates efficiently given their expertise in inverse theory. It differs from sampling-based, non-parametric variational methods in that the probabilistic solution – the posterior or post-inversion probability distribution function that describes uncertainty in the model parameters of interest – is parametrised by explicit mathematical expressions. These expressions allow for the efficient storage and transfer, and for the evaluation of the posterior probability density for any set of parameter values. We demonstrate how to use the package to solve a set of problems, including tomographic imaging using travel time data, full waveform inversion, surface wave dispersion inversion, and vertical electrical sounding. We provide built-in forward functions to simulate first arrival travel times and full acoustic waveform data (in two spatial dimensions), and external forward functions can be incorporated into the package easily. We also demonstrate how to change prior information efficiently post-inversion, using the method of variational prior replacement. Contributions from the community are welcome, to make the package more broadly applicable.