Geophysical Journal International
ISSN / EISSN : 0956-540X / 1365-246X
Published by: Oxford University Press (OUP) (10.1093)
Total articles ≅ 17,012
Latest articles in this journal
Geophysical Journal International; doi:10.1093/gji/ggab269
Simulating poroelastic waves in large-scale 3D problems having porous media coupled with elastic solids and fluids is computationally challenging for traditional methods. It is well established that the spectral element method (SEM) is more effective than the traditional methods like the finite element method (FEM) when dealing with complex geophysical problems, for its high-order accuracy with exponential convergence. However, at present, little research has been done for SEM in the frequency domain, which will be more efficient than the time-domain SEM for narrowband simulations with multiple sources, material dispersion and attenuation. Herein, we systematically develop a+ SEM in the frequency domain to simulate coupled poroelastic, elastic, and acoustic waves in anisotropic (i.e., porosity, permeability, and elastic coefficients with anisotropy), heterogeneous, and lossy media. Furthermore, we completely remove the dimension inconsistency between the displacement field and the pressure in porous media to reduce the condition number of the system matrix by around 16 orders of magnitude while maintaining the symmetry of the system matrix. To solve the multiphysics coupling problems, we apply different coupling conditions to different interface types, and use basis functions to discretize the corresponding governing equations. Numerical examples show that the proposed SEM can obtain higher accuracy with much fewer unknowns compared with the FEM and has the capacity to solve the large-scale real coupling problems.
Geophysical Journal International; doi:10.1093/gji/ggab270
We propose methods to efficiently explore the generalised nullspace of (nonlinear) inverse problems, defined as the set of plausible models that explain observations within some misfit tolerance. Owing to the random nature of observational errors, the generalised nullspace is an inherently probabilistic entity, described by a joint probability density of tolerance values and model parameters. Our exploration methods rest on the construction of artificial Hamiltonian systems, where models are treated as high-dimensional particles moving along a trajectory through model space. In the special case where the distribution of misfit tolerances is Gaussian, the methods are identical to standard Hamiltonian Monte Carlo (HMC), revealing that its apparently meaningless momentum variable plays the intuitive role of a directional tolerance. Its direction points from the current towards a new acceptable model, and its magnitude is the corresponding misfit increase. We address the fundamental problem of producing independent plausible models within a high-dimensional generalised nullspace by autotuning the mass matrix of the Hamiltonian system. The approach rests on a factorised and sequentially preconditioned version of the L-BFGS method, which produces local Hessian approximations for use as a near-optimal mass matrix. An adaptive time stepping algorithm for the numerical solution of Hamilton’s equations ensures both stability and reasonable acceptance rates of the generalised nullspace sampler. In addition to the basic method, we propose variations of it, where autotuning focuses either on the diagonal elements of the mass matrix or on the macroscopic (long-range) properties of the generalised nullspace distribution. We quantify the performance of our methods in a series of numerical experiments, involving analytical, high-dimensional, multi-modal test functions. These are designed to mimic realistic inverse problems, where sensitivity to different model parameters varies widely, and where parameters tend to be correlated. The tests indicate that the effective sample size may increase by orders of magnitude when autotuning is employed. Finally, we present a proof of principle of generalised nullspace exploration in visco-elastic full-waveform inversion. In this context, we demonstrate (1) the quantification of inter- and intra-parameter trade-offs, (2) the flexibility to change model parameterisation a posteriori, for instance, to adapt averaging length scales, (3) the ability to perform de-homogenisation to retrieve plausible sub-wavelength models, and (4) the extraction of a manageable number of alternative models, potentially located in distinct local minima of the misfit functional.
Geophysical Journal International; doi:10.1093/gji/ggab263
Observations of the ψ1 earth tide yield valuable insights into the earth’s free core nutation, especially if the effects of the ψ1 ocean tide can be removed. The ocean tide is extremely small, with amplitudes rarely more than a few millimeters, and developing an accurate model is challenging. Direct observations are inadequate to support a global model. The alternative—numerical simulation—must account for a multitude of possible effects. The ocean tide is forced by the gravitational tidal potential, by pressure loading from atmospheric tides, by seasonal modulation of the nearby K1 constituent, and possibly by nonlinear interactions among several other constituents. Here we construct a model of the ψ1 ocean tide which accounts for (or attempts to bound) each of these effects. The radiational component (from atmospheric pressure loading), although relatively small, is complicated by the presence of non-tidal atmospheric variability in the diurnal band. The ocean’s response is dynamic, but there is also high-wavenumber pressure forcing with a near-isostatic response. A general circulation model, forced by both winds and the tidal potential, suggests that annual variability in K1 leads to pronounced ψ1 amplitudes in some marginal seas, especially in the western Pacific.
Geophysical Journal International; doi:10.1093/gji/ggab267
Recent developments in sensor technology have allowed for the low-cost deployment of dense seismic arrays which continuously record the seismic ambient noise field. Generally, the extraction of body waves from the seismic ambient noise field is more challenging than for surface waves, due to the dominant amplitude of the latter. In this study, we work with data from a dense seismic array deployment in the Dehdasht area, southwestern Iran. We show that by using the polarization properties of seismic waves, we can simultaneously retrieve both high frequency body waves and surface waves from the cross-correlated noise field. As a by-product of this study, we also observe high energy spurious artefacts, particularly those associated with the direct P-phases. Numerical simulations show that these artefacts are a consequence of an uneven distribution of noise sources, and are difficult to suppress during the simultaneous retrieval of body and surface waves.
Published: 10 July 2021
Geophysical Journal International; doi:10.1093/gji/ggab265
Information on fault zone structure is essential for our understanding of earthquake mechanics, continental deformation and seismic hazard. We use the scattered seismic wavefield to study the subsurface structure of the North-Anatolian Fault Zone (NAFZ) in the region of the 1999 İzmit and Düzce ruptures using data from an 18-month dense deployment of seismometers with a nominal station spacing of 7 km. Using the forward- and back-scattered energy that follows the direct P-wave arrival from teleseismic earthquakes, we apply a scattered wave inversion approach and are able to resolve changes in lithospheric structure on a scale of 10 km or less in an area of about 130 km by 100 km across the NAFZ. We find several crustal interfaces that are laterally incoherent beneath the surface strands of the NAFZ and evidence for contrasting crustal structures either side of the NAFZ, consistent with the presence of juxtaposed crustal blocks and ancient suture zones. Although the two strands of the NAFZ in the study region strike roughly east-west, we detect strong variations in structure both north-south, across boundaries of the major blocks, and east-west, parallel to the strike of the NAFZ. The surface expression of the two strands of the NAFZ is coincident with changes on main interfaces and interface terminations throughout the crust and into the upper mantle in the tomographic sections. We show that a dense passive network of seismometers is able to capture information from the scattered seismic wavefield and, using a tomographic approach, to resolve the fine scale structure of crust and lithospheric mantle even in geologically complex regions. Our results show that major shear zones exist beneath the NAFZ throughout the crust and into the lithospheric mantle, suggesting a strong coupling of strain at these depths.
Published: 10 July 2021
Geophysical Journal International; doi:10.1093/gji/ggab268
The publisher has not yet granted permission to display this abstract.
Geophysical Journal International; doi:10.1093/gji/ggab264
We developed a three-dimensional (3D) forward modelling code, which simulates controlled-source electromagnetic problems in frequency domain using edge-based finite elements and a total electric field approach. To evaluate electromagnetic data acquired across complex subsurface structures, software performing accurate 3D modelling is required, especially for incorporation in inversion approaches. Our modelling code aims at finding a good compromise between the necessary solution accuracy at the points of interest and the general problem size by using a goal-oriented mesh refinement strategy designed for models of variable electric conductivity and magnetic permeability. To formulate an improved error estimator suitable for controlled-source electromagnetic problems, we developed literature approaches of mesh refinement further targeting three aspects. First, to generate a roughly homogeneously fine mesh discretisation around all receiver sites, our new error estimator weights the adjoint source term by the approximate decay of the electric field with increasing distance from the primal source using the expression for a homogeneous half-space. This causes almost no additional computational cost. Second, the error estimator employed in the refinement approach can be optimised for models with pronounced conductivity and magnetic permeability contrasts as often encountered in e.g. mineral prospecting scenarios by optionally including terms that measure the continuity of the normal component of current flow and the tangential component of the magnetic field across interfaces of abutting elements. Third, to avoid amplitude-dependent over-refining of the mesh, we formulate our element-wise error estimators relative to the local amplitude of the electromagnetic field. In this work, we evaluate the implemented adaptive mesh refinement approach and its solution accuracy comparing our solutions for simple one-dimensional (1D) models and a model with 3D anomalies to semi-analytic 1D solutions and a second-order finite-element code, respectively. Furthermore, a feasibility study for controlled-source electromagnetic measurements across ferrous mineral deposits is conducted. The numerical experiments demonstrate that our new refinement procedure generates problem-specific finite-element meshes and yields accurate solutions for both simple synthetic models and realistic survey scenarios. Especially for the latter, characteristics of our code such as the possibility of modelling extended sources as well as including arbitrary receiver distributions and detailed subsurface anomalies, are beneficial.
Geophysical Journal International; doi:10.1093/gji/ggab262
First-arrival traveltime tomography is one of the most used velocity model building techniques especially in sparse wide-angle acquisitions for deep crustal seismic imaging cases. Relying on the inversion of a picked attribute, the absolute traveltimes, the approach is ill-posed in terms of non-uniqueness of the solution. The latter is remedied by proper regularization or the introduction of prior information. Indeed, since traveltime kernels are vulnerable to the velocity-depth ambiguity, the inversion is stabilized by the introduction of complementary data like reflections and explicit reflectors in the velocity models. Here, we propose to supplement first-arrival traveltimes by their slopes, in other words the horizontal component of the slowness vectors at the sources and/or receivers. Slopes are a crucial attribute in state of the art scattering-based or reflection-based tomographic methods like slope tomography or wavefront tomography where the differential information is needed in order to locate the scattering events position or to parametrize the wavefront. The optional but valuable injection of slopes as an objective measure in first-arrival traveltime tomography stabilizes the problem by constraining the emergence angle or in turn implicitly the turning point depth of the rays. We explain why slopes have a tremendous added value in such a tomographic problem and highlight its remedial effect in cases where the medium is unevenly illuminated. We also show that the contribution of slopes become even more significant when the acquisition is sparse as it is generally the case with ocean-bottom seismometer surveys. The inferred models from such an extended time-attributes tomography will be used as initial guesses in a full-waveform inversion workflow context. The proposed strategy is benchmarked in 2D media against a dip section of the SEG/EAGE overthrust model and then followed by a revisit of ocean bottom seismometers data from the eastern-Nankai subduction margin as a real deep crustal case study.
Geophysical Journal International; doi:10.1093/gji/ggab261
In this study, we focus on improved constraint of the glacial isostatic adjustment (GIA) signal at present-day, and its role as a contributor to present-day sea-level budgets. The main study area extends from the coastal regions of northwestern Europe to northern Europe. Both Holocene relative sea level (RSL) data as well as vertical land motion (VLM) data are incorporated as constraints in a semi-empirical GIA model. 71 geological rates of GIA-driven RSL change are inferred from Holocene proxy data and 108 rates of vertical land motion from GNSS provide an additional measure of regional GIA deformation. Within the study area, the geological RSL data complement the spatial gaps of the VLM data and vice versa. Both datasets are inverted in a semi-empirical GIA model to yield updated estimates of regional present-day GIA deformations. A regional validation using tide gauges is presented for the North Sea, where the GIA signal may be complicated by lateral variations in Earth structure and existing predictions of regional and global GIA models show discrepancies. The model validation in the North Sea region suggests that geological data are needed to fit independent estimates of GIA-related RSL change inferred from tide gauge rates, indicating that geological rates from Holocene data do provide an important additional constraint for data-driven approaches to GIA estimation.
Geophysical Journal International; doi:10.1093/gji/ggab259
During earthquakes, structural damage is often related to soil conditions. Following the 01 April 2014 Mw 8.1 Iquique earthquake in Northern Chile, damage to infrastructure was reported in the cities of Iquique and Alto Hospicio. In this study, we investigate the causes of site amplification in the region by numerically analyzing the effects of topography and basins on observed waveforms in the frequency range 0.1–3.5 Hz using the spectral element method. We show that topography produces changes in the amplitude of the seismic waves (amplification factors up to 2.2 in the frequency range 0.1–3.5 Hz) recorded by stations located in steep areas such as the ca. 1 km-high coastal scarp, a remarkable geomorphological feature that runs north–south, that is, parallel to the coast and the trench. The modeling also shows that secondary waves—probably related to reflections from the coastal scarp—propagate inland and offshore, augmenting the duration of the ground motion and the energy of the waveforms by up to a factor of three. Additionally, we find that, as expected, basins have a considerable effect on ground motion amplification at stations located within basins and in the surrounding areas. This can be attributed to the generation of multiple reflected waves in the basins, which increase both the amplitude and the duration of the ground motion, with an amplification factor of up to 3.9 for frequencies between 1.0 and 2.0 Hz. Comparisons between real and synthetic seismic waveforms accounting for the effects of topography and of basins show a good agreement in the frequency range between 0.1 and 0.5 Hz. However, for higher frequencies, the fit progressively deteriorates, especially for stations located in or near to areas of steep topography, basin areas, or sites with superficial soft sediments. The poor data misfit at high frequencies is most likely due to the effects of shallow, small-scale 3D velocity heterogeneity, which is not yet resolved in seismic images of our study region.