Natural Sciences

Two faces of the Atlantic seafloor as heard by the UPFLOW array. (a) Station UP16 on 28 October 2021: the vertical-component waveform captures a local earthquake (E) flanked by rhythmic pulses from fin whale calls (arrows), their ~20 Hz signature clearly visible in the spectrogram below. (b) Station UP05 on 7 December 2021: superimposed waveforms from a teleseismic Mw 5.5 doublet at the Blanco transform fault off Oregon, bandpass-filtered to 14-60 s. The spectrograms reveal how seismic energy from opposite ends of the signal spectrum - biological at high frequencies, tectonic at long periods - coexists in the same ocean-bottom records, illustrating both the scientific reach and the interdisciplinary potential of the largest passive seafloor seismic experiment ever conducted in the Atlantic. Abstract Imaging deep mantle flow is essential for understanding global mantle dynamics and linking Earth’s interior to the surface. However, plume-like upwellings from the deep mantle remain poorly understood. As part of the UPFLOW project (2021-2027) funded by the European Research Council, we conducted the largest passive seafloor seismic experiment to date in the Atlantic Ocean, focusing on the Azores-Madeira-Canary Islands region, a unique setting with multiple unresolved upwellings. The project involved 11 institutions from five countries (UK, Portugal, Germany, Ireland and Spain). The deployment and recovery of ocean-bottom seismometers (OBSs) were conducted aboard the Portuguese Institute for Sea and Atmosphere (IPMA) research vessel Mário Ruivo. The data reveal a wealth of seismic signals, enabling detailed imaging of mantle upwellings and opening avenues for unexpected interdisciplinary research in marine biology and oceanography. The project also highlighted the value of outreach and education while at sea. ...

Predicting crop yields under changing climate conditions requires understanding how individual plant components develop over time, but manually measuring leaves, flowers, and pods across thousands of high-resolution images is impractical. This study adapts MapReader, originally developed for analyzing historical maps, to automatically segment and classify plant structures in whole oilseed rape (Brassica napus) images. Panel A shows the original plant image, while panels B-D compare three modeling approaches: 6-label multi-class classification (B), chain of binary classifiers (C), and the top-performing combined approach (D) that first separates plant from background, then classifies five plant structures (branches in light yellow, pods in orange, leaves in red, buds in magenta, flowers in purple). The combined approach achieved macro-averaged F1-score of 88.50 and weighted F1-score of 97.71, matching MapReader's performance on historical maps. This interdisciplinary transfer demonstrates how computer vision methods can cross domains from digital humanities to agricultural science, enabling automated phenotyping that could help ensure future food security by integrating genetic and environmental factors into crop yield models. Abstract Introduction: Recent advancements in sensor technologies have enabled collection of many large, high-resolution plant images datasets that could be used to non-destructively explore the relationships between genetics, environment and management factors on phenotype or the physical traits exhibited by plants. The phenotype data captured in these datasets could then be integrated into models of plant development and crop yield to more accurately predict how plants may grow as a result of changing management practices and climate conditions, better ensuring future food security. However, automated methods capable of reliably and efficiently extracting meaningful measurements of individual plant components (e.g. leaves, flowers, pods) from imagery of whole plants are currently lacking. In this study, we explore interdisciplinary application of MapReader, a computer vision pipeline for annotating and classifying patches of larger images that was originally developed for semantic exploration of historical maps, to time-series images of whole oilseed rape (Brassica napus) plants. ...

Thousands of small meteoroids enter Earth's atmosphere annually, but most go undetected because we rely on chance human observations or cameras pointed in the right direction. This study demonstrates a different approach: using seismometers and infrasound arrays to detect a 40 cm meteoroid that exploded unseen and unheard over the North Atlantic in June 2022. Panel (a) shows the initial source location derived from 26 seismometers and two infrasound arrays using a simplified straight-ray model, with concentric contours representing predicted arrival times converging approximately 60 km northeast of São Miguel Island in the Azores. Panel (b) presents a more sophisticated 3-D ray-tracing analysis through a realistic atmospheric model, with raypaths colored by travel time, confirming the terminal blast at 40 km altitude. The location perfectly matches an unidentified flash captured by the GOES-16 satellite's lightning detector, providing independent confirmation. This geophysical detection method, combined with satellite observations, offers a systematic way to monitor the thousands of small near-Earth objects that would otherwise remain invisible, reducing observational biases and improving our understanding of meteoroid sources and atmospheric entry processes. Abstract Small meteoroids that enter Earth’s atmosphere often go unnoticed because their detection and characterization rely on human observations, introducing observational biases in space and time. Acoustic shockwaves from meteoroid ablation convert to infrasound and seismic energy, enabling fireball detection using seismoacoustic methods. We analyzed an unreported fireball in 2022 near the Azores, recorded by 26 seismometers and two infrasound arrays. Through polarization analyses, array methods, and 3-D ray-tracing, we determined that the terminal blast occurred at 40 km altitude, approximately 60 km NE of São Miguel Island. This location matches an unidentified flash captured by a lightning detector aboard the GOES-16 satellite. The estimated kinetic energy is approximately 10⁻³ kT TNT equivalent, suggesting a 10⁻¹ m object diameter, thousands of which enter the atmosphere annually. Our results demonstrate how geophysical methods, in tandem with satellite data, can significantly improve the observational completeness of meteoroids, advancing our understanding of their sources and entry processes. ...

A fundamental assumption in reconstructing past plate motions from subducted slabs is that slabs sink vertically, marking where ancient trenches once existed. This paper reveals a dramatic exception: the Indian slab migrated laterally approximately 1,000 km northward through the mantle, pulled along by subduction of the neighboring Australian slab. By comparing tomographically imaged mantle structure with plate kinematic constraints, we show that since collision onset (45-40 Ma), the India-Asia collision zone itself moved northward as the coupled Indian-Australian plate system evolved. The schematic illustrates the process: Indian slab (green) and Australian slab (purple) separated by the extinct Wharton ridge, with the Indian slab becoming overturned and fragmented during migration. This finding has major implications for slab-based plate reconstructions worldwide: slabs in the same plate network don't sink independently but influence each other's trajectories, requiring collective interpretation to avoid potentially significant reconstruction errors. Abstract Distributions of slabs within Earth’s mantle are increasingly used to reconstruct past subduction zones, based on first-order assumptions that slabs sink vertically after slab break-off, and thus delineate paleo-trench locations. Non-vertical slab motions, which occur prior to break-off, represent a potentially significant source of error for slab-based plate reconstructions, but are poorly understood. We constrain lateral migration of the Indian slab and overlying India-Asia collision zone by comparing tomographically imaged mantle structure with plate-kinematic constraints. Following coupling of the Indian and Australian plates at the onset of collision, ∼1,000 km lateral migration of the Indian slab was driven by vertical subduction of the Australian slab. The sinking behaviors of individual slabs do not evolve in isolation, but instead influence, or are influenced by, other slabs in the same plate network. Hence, lateral slab migrations may be determined by interpreting the sinking behavior of slabs collectively, and with respect to plate kinematics. ...

Since 1971, mantle plumes have been envisioned as thin, vertical pipes of hot rock rising from the core-mantle boundary to feed volcanic hotspots like Réunion and Kerguelen. But imaging these conduits has proved frustratingly difficult because most lie beneath uninstrumented oceans. Using seismic waves that extensively sample the deepest mantle, we reveal that Indo-African upwellings are not simple vertical pipes but rather arranged in a tree-like structure: a central trunk below approximately 1,500 km depth with three tilting branches extending outward and upward toward different hotspots. This figure traces the evolution of plume geometry concepts, from Morgan's original thin conduits to our new tree-like model. We propose that each branch represents an alignment of rising blobs that detached in staggered sequence from a low-velocity corridor at the core-mantle boundary, like beads on a tilted string, each spawning a classical plume-head and tail upon reaching the viscosity contrast between lower and upper mantle. Abstract Mantle plumes were conceived as thin, vertical conduits in which buoyant, hot rock from the lowermost mantle rises to Earth’s surface, manifesting as hotspot-type volcanism far from plate boundaries. Spatially correlated with hotspots are two vast provinces of slow seismic wave propagation in the lowermost mantle, probably representing the heat reservoirs that feed plumes. Imaging plume conduits has proved difficult because most are located beneath the non-instrumented oceans, and they may be thin. Here we combine new seismological datasets to resolve mantle upwelling across all depths and length scales, centred on Africa and the Indian and Southern oceans. Using seismic waves that sample the deepest mantle extensively, we show that mantle upwellings are arranged in a tree-like structure. From a central, compact trunk below ~1,500 km depth, three branches tilt outwards and up towards various Indo-Austral hotspots. We propose that each tilting branch represents an alignment of vertically rising blobs or proto-plumes, which detached in a linear staggered sequence from their underlying low-velocity corridor at the core–mantle boundary. Once a blob reaches the viscosity discontinuity between lower and upper mantle, it spawns a ‘classical’ plume-head/plume-tail sequence. ...

The modern Andes formed by eastward subduction of Pacific seafloor beneath South America, but deeper mantle structure tells a different story. This paper analyzes four distinct slab provinces beneath South America using DETOX-P1 tomography (built from approximately 665,000 cross-correlation traveltimes), revealing evidence for a major tectonic reconfiguration around 85 Ma. The reconstruction shows a Cretaceous basin (cyan) that once separated the Andean margin from offshore intra-oceanic arcs. Left: Slabs rendered from 900 km depth downward, with South America in its 140 Ma position (translucent green). The basin occupies the gap between the reconstructed margin and western slabs (Trans-Andean and Argentine Basin). Right: As South America drifted westward, a stationary Trans-Andean trench consumed the basin. Following collision around 85 Ma, subduction polarity flipped to the modern eastward configuration, birthing the Andean arc we see today. This tectonic flip explains why slabs at different depths have dramatically different geometries despite all sinking essentially vertically. Abstract We analyze mantle structure under South America in the DETOX-P1 seismic tomography model, a global-scale, multifrequency inversion of teleseismic P waves. DETOX-P1 inverts the most extensive data set of broadband, waveform-based traveltime measurements to date, complemented by analyst-picked traveltimes from the ISC-EHB catalog. The mantle under South America is sampled by ∼665,000 cross-correlation traveltimes measured on 529 South American broadband stations and on 5,389 stations elsewhere. By their locations, depths, and geometries, we distinguish four high-velocity provinces under South America, interpreted as subducted lithosphere (“slabs”). The deepest (∼1,800–1,200 km depth) and shallowest (<600 km) slab provinces are observed beneath the Andean Cordillera near the continent’s northwest coast. At intermediate depths (1,200–900 km, 900–600 km), two slab provinces are observed farther east, under Brazil, Bolivia and Venezuela, with links to the Caribbean. We interpret the slabs relative to South America’s paleo-position over time, exploring the hypothesis that slabs sank essentially vertically after widening by viscous deformation in the mantle transition zone. The shallowest slab province carries the geometric imprint of the continental margin and represents ocean-beneath-continent subduction during Cenozoic times. The deepest, farthest west slab complex formed under intra-oceanic trenches during late Jurassic and Cretaceous times, far west of South America’s paleo-position adjoined to Africa. The two intermediate slab complexes record the Cretaceous transition from westward intra-oceanic subduction to eastward subduction beneath South America. This geophysical inference matches geologic records of the transition from Jura-Cretaceous, extensional “intra-arc” basins to basin inversion and onset of the modern Andean arc ∼85 Ma. ...

We present a quantitative plate reconstruction of western North America and the eastern Pacific basin (170 Ma to present) using "tomotectonic analysis" - integrating seismic tomography of subducted lithosphere in the mantle with surface geological data. This figure shows the trajectory of San Francisco relative to the lower mantle across different reconstructions. Previous models (black, blue) assumed continuous Farallon subduction beneath the continental margin. Our reconstruction (red) reveals a fundamentally different four-stage evolution: (1) stationary along the intraoceanic Alisitos trench, (2) transfer to the Farallon plate moving northeastward, (3) accretion and continued growth in the continent-hugging trench, and (4) northward translation to present latitude. By locating vanished plates in the deep mantle and pairing paleotrench locations with extinct arc volcanoes, we demonstrate that the eastern Pacific was broken into several smaller plates with simultaneous eastward and westward subduction under a vast archipelago, contrasting sharply with the long-held view of simple eastward Farallon subduction. Abstract Plate reconstructions since the breakup of Pangaea are mostly based on the preserved spreading history of ocean basins, within absolute reference frames that are constrained by a combination of age-progressive hotspot tracks and paleomagnetic data. The evolution of destructive plate margins is difficult to constrain from surface observations as much of the evidence has been subducted. Seismic tomography can directly constrain paleotrench locations by imaging subducted lithosphere in the mantle. This new evidence, combined with the geological surface record of subduction, suggests that several intraoceanic arcs existed between the Farallon Ocean and North America during late Mesozoic times—in contrast to existing quantitative models that typically show long-lived subduction of the Farallon plate beneath the continental margin. We present a continuously closing plate model for the eastern Pacific basin from 170 Ma to present, constrained using “tomotectonic analysis”—the integration of surface and subsurface data. During the Middle to Late Jurassic, we show simultaneous eastward and westward subduction of oceanic plates under an archipelago composed of Cordilleran arc terranes. As North America drifts westward, it diachronously overrides the archipelago and its arcs, beginning in the latest Jurassic. During and post-accretion, Cordilleran terranes are translated thousands of kilometers along the continental margin, as constrained by paleomagnetic evidence. Final accretions to North America occur during the Eocene, ending ~100 Myr of archipelago override. This model provides a detailed, quantitative tectonic history for the eastern Pacific domain, paving the way for tomotectonic analysis to be used in other paleo-oceanic regions. ...

The India-Asia collision built the Himalayas and Tibetan Plateau, but fundamental questions remain: when exactly did India collide with Asia? How large was Greater India before collision? What happened to the vast Tethys Ocean that once separated them? This study tackles these questions by integrating three independent lines of evidence: bedrock geology from Tibet recording ancient subduction, seismic tomography imaging subducted slabs now deep in the mantle, and plate kinematic reconstructions. The figure illustrates how slab geometries depend on trench behavior: stationary trenches (a-c) produce vertical or slightly tilted slabs, while mobile trenches (d-f) create more complex geometries. By assuming slabs sink vertically and matching their positions to paleo-trench locations, we test three competing models for Greater India's size. The analysis reveals that no single model satisfies all constraints, suggesting our understanding of suture zones and continental subduction limits needs revision. Abstract In this study, we integrate bedrock datasets, observations of subducted slabs in the mantle, and plate kinematic constraints to constrain models for the India-Asia collision and the central Tethys oceans. To accomplish this, we review: (1) the post-Triassic bedrock record of subduction in Tibet; (2) seismic tomographic imaging of subducted slabs in the mantle; (3) timing of the India-Asia collision; and (4) the pre-collisional size of Greater India. Following the assumption that slabs sink vertically through the mantle, their positions and geometries determined from seismic tomography constrain the locations and kinematics of paleo-subduction zones. Integrating this with bedrock constraints allows us to constrain post-Triassic subduction zone configurations for the central Tethys oceans. Neotethys was consumed by at least two subduction zones since the Jurassic. At the onset of the India-Asia collision at 59 ± 1 Ma, one subduction zone was active along the southern Asian continental margin at ~20°N. At that time, a second may have been active at subequatorial latitudes, but support for this from a bedrock perspective is lacking. This subduction zone configuration allows for three reconstructions for Greater India: The (1) minimum-area; (2) enlarged-area; and (3) Greater India Basin reconstructions. We integrate these reconstructions and subduction zone configurations in a plate kinematic framework to test their validity for the India-Asia collision. No single model is entirely satisfactory and each invokes assumptions that challenge accepted concepts. These include our understanding of suture zones, and the limits of continental subduction. We explore these challenges and their implications for our understanding of the India-Asia collision and continental collisions in general. ...

This paper presents the first global tomography study to invert a very large dataset of core-diffracted waves (Pdiff) at the highest possible frequencies, achieving resolution matching and exceeding global S-wave tomographies. Top: Schematic showing how ~10.7M traveltime measurements are inverted to create 3-D velocity models using regularized least-squares. Bottom: L-curve analysis quantifying the trade-off between model complexity (||m||) and data misfit (χ²). Moving along the L-curve varies regularization strength: too little creates grainy, noisy models; too much creates overly smooth models. Visual comparison at 2800 km depth shows the preferred DETOX-P2 model (framed in black, χ² = 1.23) that optimally balances detailed structure with noise rejection. Seismic tomography uses earthquake-generated waves recorded by global seismometer networks to create 3-D images of Earth's interior. Red colors show hot/low-velocity structures (e.g., mid-Atlantic ridge at ~20s), while blue colors show cold/high-velocity structures (e.g., subduction beneath North America at ~1:40). Abstract In global-scale seismic tomography, teleseismic P and PP waves mainly constrain structures in the upper two thirds of the mantle, whereas core-diffracted waves (Pdiff) constrain the lower third. This study is the first to invert a very large data set of Pdiff waves, up to the highest possible frequencies. This results in tomographic resolution matching and exceeding that of global S-wave tomographies, which have long been the models of choice for interpreting lowermost mantle structure. ...

Building a seismicity catalog from a single seismometer is challenging on Earth but becomes far more difficult on Mars, where noise characteristics, seismic velocities, and source types are poorly constrained. Before InSight's December 2018 landing, the marsquake service prepared the community with a blind test: 84 teams from 20 countries attempted to detect and characterize seismicity in one year of synthetic Martian data. This figure shows the true catalog of impacts and marsquakes (randomly distributed, centered on the landing site marked by white triangle), but only a fraction were detectable above Mars's noise level. Eleven teams submitted results combining established techniques with novel approaches tailored to single-station planetary seismology. The exercise generated over 100 pages of documentation, revealing the diversity of strategies needed when you cannot triangulate event locations from multiple stations and must instead rely on analyzing individual waveform characteristics, surface wave dispersion, and polarization. Abstract In December 2018, the National Aeronautics and Space Administration (NASA) Interior exploration using Seismic Investigations, Geodesy and Heat Transport (InSight) mission deployed a seismometer on the surface of Mars. In preparation for the data analysis, in July 2017, the marsquake service initiated a blind test in which participants were asked to detect and characterize seismicity embedded in a one Earth year long synthetic data set of continuous waveforms. Synthetic data were computed for a single station, mimicking the streams that will be available from InSight as well as the expected tectonic and impact seismicity, and noise conditions on Mars (Clinton et al., 2017). In total, 84 teams from 20 countries registered for the blind test and 11 of them submitted their results in early 2018. The collection of documentations, methods, ideas, and codes submitted by the participants exceeds 100 pages. The teams proposed well established as well as novel methods to tackle the challenging target of building a global seismicity catalog using a single station. This article summarizes the performance of the teams and highlights the most successful contributions. ...

Earth is not a perfect onion: its internal boundaries undulate, with topography on the core-mantle boundary reaching 10 km and transition zone thickness varying by hundreds of kilometers. Simulating seismic wave propagation through such realistic 3-D complexity at high frequencies pushes traditional methods to their limits. AxiSEM3D solves this through a hybrid approach combining spectral element and pseudospectral methods, parametrizing the azimuthal dimension with a locally adaptive Fourier series that adjusts resolution to match structural complexity. The efficiency gains are dramatic: two to three orders of magnitude faster than full 3-D methods (SPECFEM) at periods of 5 seconds or below, with speedup increasing at higher frequencies. Using particle relabelling transformation to honor undulating discontinuities while keeping the mesh spherical, AxiSEM3D enables 1 Hz simulations of 3-D mantle models with moderate computational resources, making previously inaccessible frequency ranges practical for routine use. Abstract We present a novel numerical method to simulate global seismic wave propagation in realistic aspherical 3-D earth models across the observable frequency band of global seismic data. Our method, named AxiSEM3D, is a hybrid of spectral element method (SEM) and pseudospectral method. It describes the azimuthal dimension of global wavefields with a substantially reduced number of degrees of freedom via a global Fourier series parametrization, of which the number of terms can be locally adapted to the inherent azimuthal complexity of the wavefields. AxiSEM3D allows for material heterogeneities, such as velocity, density, anisotropy and attenuation, as well as for finite undulations on radial discontinuities, both solid–solid and solid–fluid, and thereby a variety of aspherical Earth features such as ellipticity, surface topography, variable crustal thickness, undulating transition zone and core–mantle boundary topography. Undulating discontinuities are honoured by means of the ‘particle relabelling transformation’, so that the spectral element mesh can be kept spherical. The implementation of the particle relabelling transformation is verified by benchmark solutions against a discretized 3-D SEM, considering ellipticity, topography and bathymetry (with the ocean approximated as a hydrodynamic load) and a tomographic mantle model with an undulating transition zone. For the state-of-the-art global tomographic models with aspherical geometry but without a 3-D crust, efficiency comparisons suggest that AxiSEM3D can be two to three orders of magnitude faster than a discretized 3-D method for a seismic period at 5 s or below, with the speed-up increasing with frequency and decreasing with model complexity. We also verify AxiSEM3D for localized small-scale heterogeneities with strong perturbation strength. With reasonable computing resources, we have achieved a corner frequency of up to 1 Hz for 3-D mantle models. ...

Amplitude spectra from NASA's InSight lander on Mars, comparing three distinct seismic regimes. Quiet-section noise (blue, Sol 11) shows the instrument's baseline sensitivity, dominated by a broad resonance peak near 0.03 Hz. Wind noise (red, Sol 1) raises the broadband floor by roughly an order of magnitude across most frequencies. A candidate marsquake event (green, Sol 195) stands out between 0.02 and 0.1 Hz, exceeding both noise floors - the spectral signature that blind-test participants were trained to identify. Dashed lines mark diagnostic frequency bands used to distinguish genuine seismic signals from environmental noise, a key challenge given that Mars lacks the oceanic microseism that dominates Earth's ambient noise field. Abstract Benjamin Fernando, Kasra Hosseini and Maria Tsekhmistrenko describe how blind testing is helping to train martian seismologists. NASA’s InSight mission placed the first broadband seismometer on the surface of Mars, opening a new era of planetary seismology. But interpreting seismic signals on Mars poses unique challenges: the single-station configuration, unfamiliar noise environment, and absence of oceanic microseisms all differ fundamentally from terrestrial seismology. To prepare for these challenges, the InSight team organised a series of blind tests in which synthetic marsquake signals were injected into realistic noise and participants attempted to detect and characterise them. This article describes the blind-test exercise, the strategies developed by the Oxford team, and the lessons learned for identifying seismic events in an entirely new planetary environment. ...

Comparing seismic tomography models has traditionally required downloading multiple datasets, learning different formats, and writing custom visualization code. SubMachine addresses this by providing web-based tools for interactive exploration of more than 45 global and regional tomography models through a standard browser interface. The platform enables side-by-side model comparison, statistical analysis, and integration with complementary datasets including plate reconstructions, crustal structure, shear wave splitting, and gravity anomalies. By making these Earth models accessible without installation or specialized software, SubMachine facilitates collaborative research across the solid Earth community and supports quantitative comparison of different imaging approaches. Abstract We present SubMachine, a collection of web-based tools for the interactive visualization, analysis, and quantitative comparison of global-scale data sets of the Earth’s interior. SubMachine focuses on making regional and global-scale seismic tomography models easily accessible to the wider solid Earth community, in order to facilitate collaborative exploration. We have written software tools to visualize and explore over 30 tomography models - individually, side-by-side, or through statistical and averaging tools. SubMachine also serves various nontomographic data sets that are pertinent to the interpretation of mantle structure and complement the tomographies. These include plate reconstruction models, normal mode observations, global crustal structure, shear wave splitting, as well as geoid, marine gravity, vertical gravity gradients, and global topography in adjustable degrees of spherical harmonic resolution. By providing repository infrastructure, SubMachine encourages and supports community contributions via submission of data sets or feedback on the implemented toolkits. ...

Seismological research increasingly depends on large datasets, but retrieving data from multiple centers with different protocols and formats can consume more time than the actual science. ObspyDMT addresses this by providing a unified Python toolbox that handles the complexities automatically. A single command can query decades of global seismic data and generate summary visualizations. Top panel: The explosive growth of available waveforms since 1990, rising from thousands to millions of seismograms. Bottom panel: Automatic global seismicity map colored by earthquake depth. The tool requires no Python knowledge when used from the command line, yet can be integrated into automated workflows for routine tasks like data archiving, instrument correction, and quality control that are essential but time-consuming. Abstract We present obspyDMT, a free, open-source software toolbox for the query, retrieval, processing and management of seismological data sets, including very large, heterogeneous and/or dynamically growing ones. ObspyDMT simplifies and speeds up user interaction with data centers, in more versatile ways than existing tools. The user is shielded from the complexities of interacting with different data centers and data exchange protocols and is provided with powerful diagnostic and plotting tools to check the retrieved data and metadata. While primarily a productivity tool for research seismologists and observatories, easy-to-use syntax and plotting functionality also make obspyDMT an effective teaching aid. Written in the Python programming language, it can be used as a stand-alone command-line tool (requiring no knowledge of Python) or can be integrated as a module with other Python codes. It facilitates data archiving, preprocessing, instrument correction and quality control – routine but nontrivial tasks that can consume much user time. We describe obspyDMT’s functionality, design and technical implementation, accompanied by an overview of its use cases. As an example of a typical problem encountered in seismogram preprocessing, we show how to check for inconsistencies in response files of two example stations. We also demonstrate the fully automated request, remote computation and retrieval of synthetic seismograms from the Synthetics Engine (Syngine) web service of the Data Management Center (DMC) at the Incorporated Research Institutions for Seismology (IRIS). ...

Seismic tomography reveals fast anomalies in the mantle interpreted as ancient subducted slabs, but different tomographic models often disagree on where slabs exist. How do we know which features are real? This study addresses the question by creating vote maps across 14 global tomography models, identifying which anomalies appear consistently. The workflow shows our methodology: apply depth-dependent thresholds to each model (identifying on average 20% of any depth as potential slab), then count votes to find the most robust features. A striking pattern emerges with depth: agreement peaks between 1000-1400 km, declines to a minimum around 2000 km, then increases again below 2500 km. This trend could reflect reduced tomographic resolution in the mid-mantle, or alternatively real changes in time-dependent subduction flux or a mid-lower mantle viscosity increase. The vote maps reveal how conclusions can vary dramatically depending on which models you trust. Abstract The geoscience community is increasingly utilizing seismic tomography to interpret mantle heterogeneity and its links to past tectonic and geodynamic processes. To assess the robustness and distribution of positive seismic anomalies, inferred as subducted slabs, we create a set of vote maps for the lower mantle with 14 global P-wave or S-wave tomography models. Based on a depth-dependent threshold metric, an average of 20% of any given tomography model depth is identified as a potential slab. However, upon combining the 14 models, the most consistent positive wavespeed features are identified by an increasing vote count. An overall peak in the most robust anomalies is found between 1000–1400 km depth, followed by a decline to a minimum around 2000 km. While this trend could reflect reduced tomographic resolution in the middle mantle, we show that it may alternatively relate to real changes in the time-dependent subduction flux and/or a mid-lower mantle viscosity increase. An apparent secondary peak in agreement below 2500 km depth may reflect the degree-two lower mantle slow seismic structures. Vote maps illustrate the potential shortcomings of using a limited number or type of tomography models and slab threshold criteria. ...

The lowermost third of Earth's mantle remains poorly understood because the seismic waves that sample it best, core-diffracted phases (Pdiff), are difficult to model and severely attenuated. This dissertation tackles the challenge by assembling one of the largest body wave datasets for global tomography: 479,559 Pdiff traveltimes combined with millions of teleseismic P and PP measurements, processed automatically from approximately 2000 earthquakes across the broadband frequency range (up to 1 Hz). The resulting tomographic model reveals whole-mantle structures with unprecedented clarity. Vertical cross-sections beneath North America, Afar, Aegean, and Iceland show both downwellings (subducting Farallon, Tethyan, and Aegean slabs) and upwellings (mantle plumes beneath Iceland, Afar, and Tristan da Cunha). An adaptive inversion framework with locally-adjusted regularization captures sharp structural boundaries while avoiding artifacts, revealing subdivisions within the Large Low Shear Velocity Provinces and providing tomographic evidence for continuous plume conduits connecting the core-mantle boundary to surface hotspots. Abstract Seismic tomography is the pre-eminent tool for imaging the Earth’s interior. Since the advent of this method in the 1980’s, the internal structure of Earth has been vastly sampled and imaged at a variety of scales, and the resulting models have served as the primary means to investigate the processes driving our planet. Significant recent advances in seismic data acquisition and computing power have drastically progressed tomographic methods. Broad-band seismic waveforms can now be simulated up to the highest naturally occurring frequencies and consequently, measurement techniques can exploit seismic waves in their entire usable spectrum and in multiple frequencies. ...

Does a mantle plume feed the Réunion hotspot? Answering this requires imaging the mantle beneath the Indian Ocean, far from land-based seismometers. RHUM-RUM deployed 57 ocean bottom seismometers across 2000 × 2000 km² around La Réunion, the largest joint deployment of German DEPAS and French INSU instruments at the time. Photographed seconds before deployment, the instruments embody different engineering philosophies. Left: German LOBSTER with integrated sensor, optimizing for easier deployment and lower power consumption. Right: French LCPO2000-BBOBS with mechanically separated sensor (green sphere), prioritizing noise reduction at long periods (greater than 10 seconds) critical for deep mantle imaging. The trade-off: INSU instruments are quieter for studying mantle structure, while DEPAS instruments are more practical for large-scale deployments. During 13 months (October 2012 to November 2013), the network achieved 80% seismometer and 94% hydrophone data recovery, providing unprecedented coverage of this oceanic mantle plume candidate. Abstract RHUM-RUM is a German-French seismological experiment based on the sea floor surrounding the island of La Réunion, western Indian Ocean (Barruol and Sigloch, 2013). Its primary objective is to clarify the presence or absence of a mantle plume beneath the Reunion volcanic hotspot. RHUM-RUM’s central component is a 13-month deployment (October 2012 to November 2013) of 57 broadband ocean bottom seismometers (OBS) and hydrophones over an area of 2000 × 2000 km² surrounding the hotspot. The array contained 48 wideband OBS from the German DEPAS pool and 9 broadband OBS from the French INSU pool. It is the largest deployment of DEPAS and INSU OBS so far, and the first joint experiment. ...

Core-diffracted P waves (Pdiff) sample the lowermost mantle extensively but are rarely used in tomography due to modeling challenges and severe attenuation. This study processes 479,559 multifrequency traveltime measurements from 1,857 earthquakes across eight frequency bands (30.0 to 2.7 s period), making these phases usable for global waveform tomography. The maps show global coverage of core-grazing ray paths, with colors indicating sampling density. Measurement success rates decrease from 40-60% at 80° epicentral distance to only 5-10% at 140° due to attenuation, yet the resulting dataset successfully retrieves major lowermost mantle structures including the Large Low Shear Velocity Provinces beneath Africa and the Pacific, Ultra-Low Velocity Zones, and subducted slab accumulations. Abstract The lower third of the mantle is sampled extensively by body waves that diffract around the earth’s core (Pdiff and Sdiff phases), which could deliver highly resolved tomographic images of this poorly understood region. But core-diffracted waves - especially Pdiff waves - are not often used in tomography because they are difficult to model adequately. Our aim is to make core-diffracted body waves usable for global waveform tomography, across their entire frequency range. Here we present the data processing part of this effort. A method is demonstrated that routinely calculates finite-frequency traveltimes of Pdiff waves by cross-correlating large quantities of waveform data with synthetic seismograms, in frequency passbands ranging from 30.0 to 2.7 s dominant period. Green’s functions for 1857 earthquakes, typically comprising thousands of seismograms, are calculated by theoretically exact wave propagation through a spherically symmetric earth model, up to 1 Hz dominant period. Out of 418 226 candidates, 165 651 (39.6 per cent) source–receiver pairs yielded at least one successful passband measurement of a Pdiff traveltime anomaly, for a total of 479 559 traveltimes in the eight passbands considered. Measurements of teleseismic P waves yielded 448 178 usable source–receiver paths from 613 057 candidates (73.1 per cent success rate), for a total of 2 306 755 usable teleseismic dT in eight passbands. Observed and predicted characteristics of Pdiff traveltimes are discussed and compared to teleseismic P for this very large data set. Pdiff measurements are noise-limited due to severe wave attenuation with epicentral distance and frequency. Measurement success drops from 40–60 per cent at 80° distance, to 5–10 per cent at 140°. Frequency has a 2–3 times stronger influence on measurement success for Pdiff than for P. The fewest usable dT measurements are obtained in the microseismic noise band, whereas the fewest usable teleseismic P measurements occur at the highest frequencies. dT anomalies are larger for Pdiff than for P, and frequency dependence of dT due to 3-D heterogeneity (rather than just diffraction) is larger for Pdiff as well. Projecting the Pdiff traveltime anomalies on their core-grazing segments, we retrieve well-known, large-scale structural heterogeneities of the lowermost mantle, such as the two Large Low Shear Velocity Provinces, an Ultra-Low Velocity Zone west of Hawaii, and subducted slab accumulations under East Asia and Central America. ...

Need a seismogram for any earthquake at any station on Earth? Traditionally, each calculation requires running a wave propagation simulation. Instaseis changes this by precomputing and storing Green's functions in a database that enables extraction of arbitrary seismograms in milliseconds. The efficiency is remarkable: generating a complete Instaseis database costs approximately half the computational time of computing seismograms for just a single source using traditional methods. This figure shows CPU hours required to generate full databases with 1-hour seismograms for Earth and Mars using two time integration schemes. By storing basis coefficients of Lagrange polynomials rather than raw waveforms, Instaseis achieves 4th order spatial accuracy while exactly honoring velocity discontinuities like the core-mantle boundary. This transforms workflows that previously required supercomputer access into laptop-scale computations. Abstract We present a new method and implementation (Instaseis) to store global Green’s functions in a database which allows for near-instantaneous (on the order of milliseconds) extraction of arbitrary seismograms. Using the axisymmetric spectral element method (AxiSEM), the generation of these databases, based on reciprocity of the Green’s functions, is very efficient and is approximately half as expensive as a single AxiSEM forward run. Thus, this enables the computation of full databases at half the cost of the computation of seismograms for a single source in the previous scheme and allows to compute databases at the highest frequencies globally observed. By storing the basis coefficients of the numerical scheme (Lagrange polynomials), the Green’s functions are 4th order accurate in space and the spatial discretization respects discontinuities in the velocity model exactly. High-order temporal interpolation using Lanczos resampling allows to retrieve seismograms at any sampling rate. AxiSEM is easily adaptable to arbitrary spherically symmetric models of Earth as well as other planets. In this paper, we present the basic rationale and details of the method as well as benchmarks and illustrate a variety of applications. ...

Computing how seismic waves propagate through Earth's full 3-D structure at high frequencies typically requires supercomputers running for days or weeks. AxiSEM achieves the same accuracy in a fraction of the time by exploiting a key simplification: for spherically symmetric Earth models, the azimuthal dimension can be computed analytically rather than numerically. This reduces the computational domain from 3-D to 2-D while maintaining full 3-D accuracy in the output wavefields. The figure shows a 3-D wavefield simulation from an earthquake in Italy, computed with AxiSEM at frequencies across the observable seismic band. This efficiency breakthrough enables applications previously impractical, from computing millions of synthetic seismograms for tomographic inversions to generating databases for near-instantaneous retrieval of seismograms from any source-receiver combination. Seismic wave propagation in a spherically symmetric Earth model computed using AxiSEM. Warm colors (red/yellow) show P-waves, while cold colors (blue/green) show S-waves. Abstract We present a methodology to compute 3-D global seismic wavefields for realistic earthquake sources in visco-elastic anisotropic media, covering applications across the observable seismic frequency band with moderate computational resources. This is accommodated by mandating axisymmetric background models that allow for a multipole expansion such that only a 2-D computational domain is needed, whereas the azimuthal third dimension is computed analytically on the fly. This dimensional collapse opens doors for storing space–time wavefields on disk that can be used to compute Fréchet sensitivity kernels for waveform tomography. We use the corresponding publicly available AxiSEM (www.axisem.info) open-source spectral-element code, demonstrate its excellent scalability on supercomputers, a diverse range of applications ranging from normal modes to small-scale lowermost mantle structures, tomographic models, and comparison with observed data, and discuss further avenues to pursue with this methodology. ...

As seismological data centers exploded with new waveform data in the 2010s, researchers faced a growing challenge: downloading, homogenizing, and quality-controlling data from multiple centers with different interfaces and formats consumed more time than the actual science. ObsPyLoad addresses this data avalanche with a fully automated solution that queries metadata holdings and retrieves seismograms from multiple data centers simultaneously - a major advantage over individual center tools. This schematic shows the elegant program flow: from initial configuration through metadata queries to IRIS and ORFEUS data centers, followed by parallel waveform downloads with built-in quality control and retry logic. A simple command-line call without parameters downloads event-based data for all earthquakes from the past 30 days, while extensive options allow customization for geographic regions, time windows, magnitude thresholds, and update modes. By handling the tedious infrastructure work automatically, ObsPyLoad lets seismologists focus on scientific questions rather than data wrangling. Abstract We confront the data avalanche: the amount of waveform data available from seismological data centers has been growing enormously over the past few years. This is a highly welcome development from a scientific point of view, but the time and effort spent on identification, retrieval, and quality control of subsets of these data may quickly exceed tolerable limits for an individual researcher. ...