Waveform Tomography Research Articles

WUS324: Multiscale Full Waveform Inversion Approaching Convergence Improves Waveform Fits While Imaging Seismic Structure of the Western United States

AbstractWe report a new model of radially anisotropic crustal and upper mantle structure of the western United States (WUS324) obtained from full waveform inversion of earthquake data. We ran three multiscale inversion stages beyond model WUS256 (Rodgers et al., 2022, https://doi.org/10.1029/2022jb024549) allowing them to approach convergence to fit a larger data set to a shorter minimum period of 16 s. WUS324 is based on 324 total iterations from its starting model, significantly more (16 times) than previous studies. Waveform misfit reductions are 66%–70% for both the inversion data and an independent validation data set providing confidence in the predictive power of the model. WUS324 provides much better fits and reveals shear wavespeed, vS, structure of this large region with more detail than previous waveform tomography models. We show representative images demonstrating the resolution of diverse seismic structure across this highly heterogeneous region including oceanic lithosphere, subducting slabs and continental magmatism.

Sound speed imaging of cortical bone and soft tissue based on frequency-domain ultrasound waveform tomography

Abstract As one of the ultrasound computed tomography (UCT) algorithms, full-waveform inversion (FWI) can recover clear and detailed internal structures of bones by iteratively minimizing the misfit between observed and numerical data. However, time domain full-waveform inversion (TDFWI) often lacks sufficient low-frequency information, and the computation of full-time data imposes a significant computational burden. Besides, there are few studies on whole-medium reconstruction that jointly invert the bone and surrounding soft tissues. In this study, an FWI-based frequency-domain UCT (FD-UCT) is utilized for simultaneously imaging the sound speed of bones and soft tissues, which facilitates the inversion efficiency while ensuring the existence of low-frequency information. Simulated studies show that the FD-UCT can both reconstruct bones with large acoustic impedance and distinguish subtle impedance differences within soft tissues to reveal lesions.

Improved Earthquake Source Parameters with 3D Wavespeed Models in California and Nevada

Abstract Seismic tomography harnesses earthquake data to explore the inaccessible structure of the Earth. Adjoint waveform tomography (AWT), a method of seismic tomography, updates the tomographic model by optimizing the fit between observed earthquake data and synthetic waveforms. The synthetic data are calculated by solving the wave equation through a given 3D model. An important requirement to calculating synthetics is the source information (location, centroid time, depth, and moment tensor). Errors in source information affect the quality of the synthetics produced, which in turn can limit how structure can be inferred in the AWT workflow. To test the effect of updating source information, we used MTTime (Chiang, 2020), a time-domain full-waveform moment tensor inversion code, to calculate the moment tensors and depths of 118 earthquakes that occurred in California and Nevada over a 20-yr period. We calculated 3D Green’s functions using a 3D seismic wavespeed model of California and Nevada (Doody et al., 2023b). We show that the inverted solutions provide better waveform fits than the Global Centroid Moment Tensor catalog and increase usable, well-correlated data by up to 7%. Therefore, we argue that recalculating source parameters should be considered in AWT workflows, particularly for smaller magnitude events (Mw<5.0).

Error Propagation and Control in 2D and 3D Hybrid Seismic Wave Simulations for Box Tomography

ABSTRACT To enhance the local resolution of global waveform tomography models, particularly in areas of interest within the Earth’s deep structures, a higher resolution localized tomography approach (referred to as “box tomography”) is crucial for a more detailed understanding of the Earth’s internal structure and geodynamics. Because the small-scale features targeted by box tomography are finer than those in global reference models, distinct spatial meshes are necessary for global and local (hybrid) forward simulations. Within the spectral element method (SEM) framework, we employ the intrinsic Lagrangian spatial interpolation to compute and store hybrid inputs (displacement/potential) in the global numerical simulation. These hybrid inputs are subsequently imposed into the localized domain during the iterative box tomography. However, inaccurate spatial Lagrange interpolation can lead to imprecise hybrid inputs, and this error can propagate from the global simulation to the hybrid simulation. It is essential to quantitatively analyze this error propagation and control it to ensure the credibility of box tomography. We introduce a unique spatial window function into the conventional “direct discrete differentiation” hybrid method. When the local mesh and structure align with those in the global simulation, the synthetic hybrid waveforms match the global ones, serving as a reference for quantitatively assessing error propagation stemming from changes in the local spatial mesh during hybrid simulation. Significantly, the relative waveform error arising due to spatial Lagrange interpolation is around 5% when employing the traditional SEM with five Gauss–Lobatto–Legendre points per minimum wavelength in the 3D global simulation through SPECFEM3D_GLOBE. Ultimately, we achieve hybrid waveforms with an accuracy of about 1.5% by increasing the spectral elements by about 1.5 times in the standard global simulation.

Adjoint Waveform Tomography for Crustal and Upper Mantle Structure of the Middle East and Southwest Asia for Improved Waveform Simulations Using Openly Available Broadband Data

ABSTRACT We present a new model of radially anisotropic seismic wavespeeds for the crust and upper mantle of a broad region of the Middle East and Southwest Asia (MESWA) derived from adjoint waveform tomography. The new model enables fully 3D simulations of complete three-component waveforms and provides improved fits that were not possible with previous models. We inverted over 32,000 waveforms from 192 earthquakes recorded by over 1000 openly available broadband seismic stations from permanent and temporary networks in the region with highly uneven coverage. Inversion iterations proceeded from the period band 50–100 s in six stages and 54 total iterations reducing the minimum period to 30 s. Our final model, MESWA, improves waveform fits compared to the starting and other models for both the data used in the inversion and an independent validation set of 66 events. Restitution tests indicate that the model resolves features in the central part of the model to depths of about 150 km. The new model reveals tectonic features imaged by other studies and methods but in a new holistic model of anisotropic shear and compressional wavespeeds (VS and VP, respectively) covering a larger domain with smaller scale length and amplified features. Examples include low crustal VS in the Tethyan belt and low mantle VS following divergent (Gulf of Aden, Red Sea) and transform (Dead Sea fault) margins of the Arabian plate. Low VS is imaged below Cenozoic volcanic centers of the Mecca–Madina–Nafud Line, Arabian Peninsula, and the Türkiye–Iran border region. Elevated VS tracks Makran subduction under southeast Iran with near vertical dip. MESWA could be used as a starting model for further improvements, say, using waveforms from in-country seismic networks that are not currently openly available and/or smaller-scale studies targeting a shorter period. The model could be used to improve earthquake hazard studies and nuclear explosion monitoring.

Topography effect on seismic waveform tomography: a quantitative study

SUMMARY In seismic tomography practices the Earth's surface is sometimes assumed to be either spherical or flat for convenience in forward modelling calculations. The effect of irregular surface topography on seismic wave propagation is thus ignored, resulting in biases in the phases and amplitudes of synthetic seismograms, which contribute to the residuals that are mapped into velocity structures in tomography inversions. In this study, we conduct a series of inversion experiments based on the adjoint waveform tomography method to quantitatively assess the topography effect on waveform-based inversion results. We first employ models with simplified topography to better highlight and quantify the topography effect. Results show that when topography effect is ignored in the forward modelling, it is mapped into velocity perturbations, leading to spurious velocity anomalies in tomography models. The strength of the spurious velocity anomaly is quasi-linearly related to locally averaged topography gradient. Our inversion experiments demonstrate that in places of strong topography variation, such as the Longmenshan Fault Zone region where the 2008 Mw 7.9 Wenchuan earthquake occurred, topography effect can lead to spurious relative velocity anomalies of up to 10 per cent, which cannot be ignored in waveform-based tomography inversions.

Joint Distributed Traveltime and Full Waveform Tomography for Enhanced Subsurface Imaging in Seismic Networks

Joint Distributed Traveltime and Full Waveform Tomography for Enhanced Subsurface Imaging in Seismic Networks

CANVAS: An Adjoint Waveform Tomography Model of California and Nevada

AbstractWe present the California‐Nevada Adjoint Simulations (CANVAS) model, an adjoint waveform tomography model of the crust and uppermost mantle of the states of California and Nevada. We used WUS256 (Rodgers et al., 2022, https://doi.org/10.1029/2022jb024549) as the starting model and iteratively decreased the minimum period of CANVAS from 30 to 12 s. CANVAS was iterated in two distinct stages: the first stage with source mechanisms from the Global Centroid Moment Tensor (GCMT) catalog and the second stage with inverted moment tensors (MT) using the CANV_WUS model (Doody et al., 2023, https://doi.org/10.1029/2023jb026463). We show that updating the MTs with 3D Green's functions improved waveform fits and azimuthal coverage of windowed data used to calculate the gradients. As for the model itself, we improved waveform fits over WUS256, particularly in the dispersed surface waves. CANVAS resolved tectonic features seen in other models and accurately defined the depth to basement of major basins, including the Central Valley and the Ventura Basin. We propose CANVAS as a starting model for crustal tomography models on smaller scales.

Full-waveform simulation of DAS records, response and cable-ground coupling

SUMMARY Over the past several years, the use of optical fibre cables as ground motion sensors has become a central topic for seismologists, with successful applications of distributed acoustic sensing (DAS) in various key fields such as seismic monitoring, structural imaging and source characterization. DAS response is a combination of both instrument response and cable-ground coupling, with the latter having a strong, spatially variable, but yet largely unquantified effect. This limits the application of a large number of staple seismological techniques (e.g. earthquake magnitude estimation, waveform tomography) that can require accurate knowledge of a signal's amplitude and frequency content. Here we present a method for accurately simulating a DAS cable and its ground coupling. The scheme is based on molecular dynamic-like particle-based numerical modelling, allowing the investigation of the effect of varying DAS-ground coupling scenarios. We start by computing the full strain field directly, for each pair of neighbouring particles in the model. We then define a virtual DAS cable, embedded within the model and formed by a single string of interconnected particles. This allows us to control all aspects of the cable-ground coupling and their properties at an effective granular level through changing the bond stiffness and bond types (e.g. non-linearity) for both the cable and the surrounding medium. Arbitrary cable geometries and heterogeneous materials can be accommodated at the desired scale of investigation. We observe that at the meter scale, the cable-ground coupling and local site effects can substantially alter the recorded signal. We find that the stiffness of the thin layer of material to which the cable is coupled has the strongest effects, selectively amplifying portions of the wave train and contributing to substantial phase delays. These differences show that cable coupling and local site effects should be considered both when designing a DAS deployment and analysing its data when either true or along-cable relative amplitudes and/or frequencies are considered. The codes developed herein for calculating full waveform DAS responses and coupling are made publicly available.

Tectonic transition revealed by upper mantle heterogeneities and anisotropy of the SE margin of the Tibetan Plateau: Insights into the Cenozoic intraplate volcanisms

The SE margin of the Tibetan Plateau is a key channel for understanding the plateau's evolution. However, the links between mantle dynamics, surface deformation, and associated intraplate volcanism remain elusive. Here, 3D radially anisotropic seismic structures of the upper mantle beneath the SE margin of the Tibetan Plateau are presented based on adjoint waveform tomography using dense arrays. The seismic image shows a prominent tectonic transition along 26°N, characterized by strong heterogeneities. In the north of 26°N, lithospheric delamination is revealed by high-velocity anomalies between the Tibetan Plateau and the Yangtze Craton at the 150–250 km depth, which may drive the early Cenozoic potassic volcanism. In the south of 26°N, asthenospheric flow is suggested by significantly low velocity and positive radial anisotropy at the uppermost mantle, which probably accounts for the late Cenozoic sodic volcanism. We propose that such upper mantle dynamics shape the surface expression of plateau uplift and intraplate volcanism. Plain language summaryThe mechanism controlling the growth and volcanism of the Tibetan Plateau remains ambiguous. We present clues from subsurface structures using earthquake seismology. We observe strong heterogeneities of velocity and anisotropy in the upper mantle of the SE margin of the Tibetan Plateau. The results suggest that lithospheric delamination and asthenospheric flow control the generation of the early Cenozoic K-rich volcanism and late Cenozoic Na-rich volcanism, respectively. Our study indicates that tectonic transition might exist in the subsurface of the SE margin of the Tibetan Plateau.

Seismic Waveform Tomography for 3D Impedance Model with Salt Structure

Conventional impedance inversion from post-stack zero-offset seismic data is usually based on the convolution model, and wave-equation based inversion is rarely used, although it is capable to precisely describe seismic wave propagation and invert impedance model with higher resolution. That is because there are more than one physical parameters involved in the conventional wave equation, making impedance inversion complicated. In this study, a one-dimensional (1D) wave equation, containing only the impedance parameter, is adopted and applied for the inversion of 1D impedance model by seismic waveform tomography. However, for a three-dimensional (3D) model, direct application of the 1D waveform tomography may lead to lateral discontinuities. Therefore, we propose to utilize a truncated Fourier series to parameterize the 3D impedance model, and then invert for the Fourier coefficients. With this parameterization scheme, the large- and small-scale components of the impedance model can be reconstructed stepwise by gradually increasing the number of Fourier coefficients. To efficiently and effectively invert the coefficients for the 3D model with salt structure, we propose a joint strategy, in which the low-frequency seismic data is used to invert for the Fourier coefficients representing the large-scale components of the model, while the high-frequency seismic data is applied to invert for the Fourier coefficients representing the small-scale components of the model. Tests on a 3D impedance model with salt structure result in models with high resolution and good spatial continuity, proving the feasibility and stability of the impedance inversion procedure.

Comparing Adjoint Waveform Tomography Models of California Using Different Starting Models

AbstractAdjoint waveform tomography (AWT) sits at the cutting edge of seismic tomography on local, regional, and global scales. However, the choice in starting model may have a significant impact on the final inversion results. In this paper, we present 3 AWT models of California that are based on different starting models. We chose three models that were inverted at different scales: SPiRaL, a global travel‐time tomography model (Simmons et al., 2021, 10.1093/gji/ggab277), CSEM_NA, a regional adjoint tomography model of North America and the North Atlantic (Krischer et al., 2018, 10.1029/2017JB015289), and WUS256, a regional adjoint tomography model of the western US (Rodgers et al., 2022, https://doi.org/10.1029/2022JB024549). We then inverted three AWT models using the same source and receiver set. We ran each model over three period bands: 30–100 s, 25–100 s, and 20–80 s. Once the iterations were finalized, we used five methods of testing model similarity in both the model and data space. We conclude that the choice of starting model has a minimal impact on long wavelength models if an appropriate multi‐scale inversion approach is used.

Correction to: Structure and evolution of the Australian plate and underlying upper mantle from waveform tomography with massive data sets

Correction to: Structure and evolution of the Australian plate and underlying upper mantle from waveform tomography with massive data sets

Structure and evolution of the Australian plate and underlying upper mantle from waveform tomography with massive data sets

SUMMARY We present a new S-wave velocity tomographic model of the upper mantle beneath the Australian Plate and its boundaries that we call Aus22. It includes azimuthal anisotropy and was constrained by waveforms from 0.9 million vertical-component seismograms, with the densest data sampling in the hemisphere centred on the Australian continent, using all available data covering this hemisphere. Waveform inversion extracted structural information from surface waves, S- and multiple S-waves and constrained S- and P-wave speeds and S-wave azimuthal anisotropy of the crust and upper mantle, down to the 660-km discontinuity. The model was validated by resolution tests and, for particular locations in Australia with notable differences from previous models, independent inter-station measurements of surface-wave phase velocities. Aus22 can be used to constrain the structure and evolution of the Australian Plate and its boundaries in fine detail at the regional scale. Thick, high-velocity (and, by inference, cold) cratonic lithosphere occupies nearly all of western and central Australia but shows substantial lateral heterogeneity. It extends up to the northern edge of the plate, where it collides with island arcs, without subducting. Diamondiferous kimberlites and lamproite deposits are underlain by cratonic lithosphere, except for the most recent diamondiferous lamproites in the King Leopold Orogen. The rugged eastern boundary of the cratonic lithosphere resolved by the model provides a lithospheric definition of the Tasman Line. Just east of the Tasman Line, an area of intermediate-thick lithosphere is observed in the southern part of the continent. The eastern part of Australia is underlain by thin, warm lithosphere, evidenced by low seismic velocities. All the sites of Cenozoic intraplate volcanism in eastern Australia are located on thin lithosphere. A low-velocity anomaly is present in the mantle transition zone (410–660 km depths) beneath the Lord Howe and Tasmanid hotspots, indicative of anomalously high temperature and consistent with a deep mantle upwelling feeding these hotspots and, possibly, also the East Australia hotspot. High seismic velocities at 200–410 km depth below New Guinea indicate the presence of slab fragments, probably linked to the subduction of the Australian Plate. High seismic velocities are observed in the transition zone below northeast Australia and indicate the presence of subducted lithospheric fragments trapped in the transition zone, possibly parts of the former northern continental margin of Australia.

Dense-array adjoint tomography reveals lithospheric delamination and asthenosphere upwelling beneath the western Yangtze Craton

The eastern Tibetan Plateau has attracted widespread attention due to its complex topography and strong seismicity. However, the mechanism controlling the growth of this margin remains enigmatic. Here, we present detailed upper mantle structures of the easternmost and northeastern Tibetan Plateau from dense-array adjoint waveform tomography. The seismic images show mushroom-shaped low-velocity zones atop at the uppermost mantle of the eastern margin of the Tibetan Plateau and slab-shaped high-velocity bodies preserved beneath both the eastern Himalayan syntaxis and the western Sichuan basin. The seismic features suggest asthenosphere upwelling and lithospheric delamination beneath the western Yangtze Craton, which might be induced by the retreat of the subducted Indian Plate. Our study shows that the mantle dynamics of the western Yangtze Craton have played an important role in the Tibetan Plateau growth and suggests that the Yangtze Craton might be reconstructed by ongoing continental collision.

Double‐Difference Adjoint Tomography of the Crust and Uppermost Mantle Beneath Alaska

AbstractWe perform adjoint waveform tomography to reveal the P‐wave velocity structure of the crust and uppermost mantle in Alaska by using common‐source double‐difference traveltime data. Our underlying forward modeling tool is a 3D seismic‐wave solver called SPECFEM3D_GLOBE. We select ∼13,000 high‐quality P‐wave arrivals from 147 earthquakes recorded by more than 600 stations. The waveforms are filtered between 15 and 8 s and used to generate about 100,000 differential traveltimes via the cross‐correlation technique. We invert these common‐source differential traveltimes for a P‐wave velocity model down to 150 km depth in Alaska. In the upper mantle, a high‐velocity zone is imaged along the Aleutian volcanic arc, which represents the Pacific plate subducting beneath the North America plate. On the eastern edge of the subducting Pacific plate, the imaged high‐velocity anomaly extends distinctly beyond the Wadati–Benioff zone, indicating an aseismic slab edge. We observe an absence of a low‐velocity mantle wedge beneath the Denali Volcanic Gap (DVG) in our tomographic model. This finding suggests that minimal melt accumulation exists beneath the DVG, which may explain the cessation of magmatism there. Our model also reveals a high‐velocity anomaly near the Wrangell Volcanic Field (WVF), suggesting the possible existence of the Wrangell slab or the high‐velocity slab edge of the Yakutat slab. A potential slab gap shown as a low‐velocity body is detected at 95–125 km depth near the WVF, which could act as a channel to transport mantle materials to feed the cluster of volcanoes in the WVF.

A Complex Mantle Plume Head Below East Africa‐Arabia Shaped by the Lithosphere‐Asthenosphere Boundary Topography

AbstractHot plumes rising from Earth's deep mantle are thought to cause uplift, rifting and large igneous province (LIP) emplacement. LIP volcanism in continents often spans tens of Ma and scatters unevenly over broad areas. This has been attributed to lateral flow of hot plume material, but observational evidence on such flow is scarce. New waveform tomography with massive data sets reveals detailed seismic velocity structure beneath the East Africa‐Arabia region, where these processes occur at present. It shows interconnected sub‐lithospheric corridors of hot, partially molten rock, fed by three mantle upwellings beneath Kenya, Afar, and Levant. The spatio‐temporal distribution of the volcanism suggests that we are witnessing an integral plume head, which morphed into a three‐pointed star by ponding and channeling within thin‐lithosphere corridors. Plate reconstructions indicate that it spread south‐to‐north since ∼45 Ma. These results suggest that complex‐shape plume heads can explain the enigmatic, scattered LIP volcanism and are, probably, an inherent feature of plume‐continent interaction.

Mitigating the effect of errors in source parameters on seismic (waveform) tomography

SUMMARYWe investigate the effect of errors in earthquake source parameters on the tomographic inverse problem and propose mitigation strategies for avoiding artefacts caused by such errors. In global catalogues, mislocation errors can amount to tens of kilometres both horizontally and vertically, while fault plane uncertainties can be of the order of tens of degrees. We conduct a perturbation study investigating both errors in the source location and in the moment tensor. Spatial perturbations of 5 km and fault plane perturbations of 5° result in measured time-shifts of the order of 0.5 to several seconds, which in five iterations lead to artefacts with amplitudes of the order of 0.5–1 per cent spanning up to several hundreds of kilometres. Larger perturbations (e.g. 20 km) lead to artefacts similar in amplitude (∼5 per cent) to the features judged to be interpretable in tomographic models. This can be further exacerbated by the cumulative effect of systematic errors. Mitigation strategies consist of removing a region around the source from the gradient and discarding traces where amplitudes are comparatively small. The most problematic type of error is a horizontal mislocation, because its imprint is not removed by such measures—discarding a ‘suspicious’ event may be the only option if no trustworthy ground truth is available. Although this study focuses on (adjoint) waveform tomography, a large part of the results are equally valid for any other type of imaging method that is based on time- and/or phase-shift measurements. The concerns and mitigation strategies presented here therefore have broader applicability.

WUS256: An Adjoint Waveform Tomography Model of the Crust and Upper Mantle of the Western United States for Improved Waveform Simulations

AbstractWe report a new model (WUS256) of radially anisotropic seismic wavespeeds of the crust and upper mantle of the western United States (WUS) obtained from adjoint waveform tomography for the purpose of improving synthetic waveform fits to observed data. WUS256 is based on inversion of over 94,000 waveforms from 72 earthquakes recorded by nearly 3,400 stations. We started with the SPiRaL global model (Simmons et al., 2021, https://doi.org/10.1093/gji/ggab277) and waveforms in the period band of 50–120 s. We followed a conservative multiscale inversion approach with eight stages and 256 total inversion iterations which enabled monotonic misfit reduction to 20‐s minimum‐period waves. WUS256 relied on time‐frequency (TF) phase misfits and a trust region limited memory Broyden–Fletcher–Goldfarb–Shanno (L‐BFGS) optimization. Hessian‐vector products were used to qualitatively assess model resolution. Results indicate that WUS256 has good coverage of the continental regions to depths of about 150 km and is able to resolve features on lateral scales of about 200 km. We quantify waveform fits by the reduction in TF and normalized amplitude difference misfits between WUS256 and the SPiRaL starting model. WUS256 significantly improves waveform fits with misfit reduction 64% for both inversion and validation data sets compared to the SPiRaL starting model and shows even better fits compared to other models. Waveform fits illustrate that WUS256 reproduces body‐waves, fundamental mode surface waves as well as late arriving dispersed and/or scattered short period surface waves. The improvement in waveform fit indicates that WUS256 can be used to reproduce path effects on regional complete waveforms and moment tensor inversions.

Propagating Seismic Waves in VTI Attenuating Media Using Fractional Viscoelastic Wave Equation

AbstractSeismic velocity and attenuation anisotropy are ubiquitous in the crust and upper mantle, significantly modulating the characteristics of seismic wave propagation in the Earth's interior. Accurate seismic wave modeling of velocity and attenuation anisotropy is essential for the understanding of wave propagation in the Earth's interior as well as constructing global and region‐scale seismic full waveform tomography. Here, we derive a decoupled fractional Laplacian (DFL) viscoelastic wave equation to characterize the Earth's frequency‐independent Q behavior in the vertical transversely isotropic (VTI) media. We verify the accuracy of the proposed viscoelastic wave equation by 2D synthetic examples; to show its applicability in crustal‐scale seismic modeling, we present an example of 3D seismic wave propagation in the realistic Salton Trough model. Through extensive numerical tests, we conclude that the proposed viscoelastic wave equation is superior in four aspects. First, the viscoelastic wave equation takes VTI anisotropy of both velocity and attenuation into account and can describe the strong direction‐dependent attenuation. Second, our derivation contains spatially independent Laplacians, and thus the proposed wave equation enjoys higher simulation accuracy for heterogeneous Q media. Third, the new viscoelastic wave equation can decouple the amplitude decay and the phase distortion, which is appealing for improving the resolution in seismic imaging and inversion. Lastly, compared to viscoelastic wave equations with time‐fractional operators, our scheme has higher computational efficiency by avoiding substantial wavefield storage.