Search

Rifted margins often show apparent along‐strike variations at the continent‐ocean transition (COT) region, typically occurring gradually over hundreds‐to‐thousands of km between the "magma‐rich" and "magma‐poor" endmembers. In contrast, comparatively smaller‐scale lateral variabilities of rifting structure and faulting style in hyper‐thinned continental crust are not yet well described. The South China Sea (SCS) was recently defined as the archetype of a new "intermediate‐type" margin with neither excess magmatism nor mantle exhumation. However, along‐strike crustal variations along this new type of margin are still little explored. Based on previously unpublished 2D grid of seismic reflection lines and overlapping 3D seismic volumes, we have investigated the crustal structure, faulting style, and the lower crustal reflectivity along ∼400 km of the mid‐northern SCS margin. Our new maps of crystalline‐crust thickness and top basement show that the SCS extension developed in two segments with contrasting tectonic styles separated by a narrow boundary. The eastern domain, resembling a wide‐rift model, displays low‐angle detachment faults and high‐velocity lower crust mainly associated with syn‐rift magmatism indicating weak crustal rheology. The neighboring western domain, resembling a narrow rift model, displays steep crustal‐scale faults indicating comparatively strong rheology. The change from wide to narrow rift—or weak to strong rheology—occurs abruptly, which appears at odds with the current conventional wisdom of mostly gradual transitions over a long distance between rift models. We speculate that the inherited heterogeneity in mantle composition and 3D melt focusing effect may drive this abrupt along‐strike variability in the mid‐northern SCS. Plain Language Summary: The South China Sea (SCS) rifted margin recently defined a type of rifted margin worldwide, namely the "intermediate‐type" model that is different from the conventional "magma‐rich" and "magma‐poor" models. This new margin type can be somewhat applied to many other rifted margins, for example, central South Atlantic margins, Gulf of Lions, Northern Red Sea, the Gulf of Aden. Understanding the crustal architecture of the SCS from a 3‐D perspective is the key to studying the lateral structural variations of the "intermediate‐type" margins in the world. However, the limitations of data coverage and resolution in previous studies led to insufficient knowledge of the 3D crustal structure in the SCS. For the first time, we provide a large‐scale high‐resolution crustal thickness map based on the interpretation in largely unpublished 2D seismic grid imaging of the whole crustal structure. Our results reveal an abrupt change in tectonic rifting style, rift‐related magmatism, and rheological behavior along the SCS margin. Two distinct tectonic domains with contrasting rifting modes are therefore defined and inferred to result from lateral variability of mantle composition and 3D melt focusing effect. Key Points: 3D crustal structure variation along the mid‐northern South China Sea based on seismic interpretation mapped for the first timeContrasting rift modes with steep crustal‐scale brittle faulting in the west versus low‐angle detachments soling out in ductile lower crust in the eastAbrupt rheological lateral changes possibly associated with inherited compositional mantle heterogeneities and 3D melt focusing [ABSTRACT FROM AUTHOR]

Seismic interpretation aims to extract quantitative and interpretable attributes from a seismic image produced using some migration method to inform characteristics of a subsurface reservoir or target of interest. Current paradigms for computing seismic attributes mostly rely on single-task algorithms. We develop an iterative, multitask machine learning method to learn and infer multiple attributes from a seismic image. This method is composed of two stages: a multitask inference stage and a multimodal, multitask refinement stage. The basic mechanism of this method is that we train a multitask inference neural network to estimate a set of attributes, including a relative geological time volume, a denoised higher-resolution seismic image and multiple fault attributes (including probability, dip and strike), from a low-resolution, noisy seismic image; then we input the inferred attributes to a multitask refinement NN to enhance the raw inference results iteratively. The two multitask neural networks are trained separately based on synthetic seismic images and associated labels generated by geological modelling. Applications of this multitask learning and inference method to synthetic and field seismic images show that our method can improve the structural consistency among output seismic attributes compared with single-task neural networks, leading to more reliable automatic interpretation and subsurface characterization. [ABSTRACT FROM AUTHOR]

The Nazca-South America subduction zone in Ecuador is characterized by a complicated along-strike geometry as the slab transitions from flat slab subduction in the south, with the Peruvian flat slab, to what has been characterized as 'normal' dipping subduction beneath central Ecuador. Plate convergence additionally changes south to north as the trench takes on a convex shape. Highly heterogeneous bathymetry at the trench, including the aseismic oceanic Carnegie Ridge (CR), and sparse intermediate-depth seismicity has led many to speculate about the behaviour of the downgoing plate at depth. In this study, we present a finite-frequency teleseismic P -wave tomography model of the northern Andes beneath Ecuador and Colombia from 90 to 1200 km depth. Our model builds on prior tomography models in South America by adding relative traveltime residuals recorded at stations in Ecuador. The complete data set is comprised of 114 096 relative traveltime residuals from 1133 stations across South America, with the added data serving to refine the morphology of the Nazca slab in the mantle beneath the northern Andes. Our tomography model shows a Nazca slab with a fragmented along-strike geometry and the first teleseismic images of several proposed slab tears in this region. At the northern edge of the Peruvian flat slab in southern Ecuador, we image a shallow tear at 95–200 km depth that appears to connect mantle flow from beneath the flat slab to the Ecuadorian Arc. Beneath central Ecuador at the latitudes of the CR, the Nazca slab is continuous into the lower mantle. Beneath southern Colombia, the Malpelo Tear breaks the Nazca slab below ∼200 km depth. [ABSTRACT FROM AUTHOR]

Deep learning has shown a considerable potential to significantly improve processing efficiency but has not yet been widely deployed to production projects of seismic signal separation such as seismic interference attenuation. The main reasons are: First, the industry has high standards for signal fidelity, which are critical for the success of subsequent seismic imaging, and deep neural network methods have not yet matched the required level; second, the network's interpretability issue has affected many geophysicists and sponsors' trust in the deep learning technique. To develop deep neural network methods towards the end of benefiting real‐world production, we first attempt to better understand their performance, especially in how they make use of local and global features of the data. A novel quantitative research of the overall network model behaviour on synthetic data is conducted. We simulate three types of coherent seismic data components in the shot domain, blend them together and then train a network to separate them. In this process, random noise, a component having only learnable local features, is selectively injected into the network's training pairs. Three network models sharing the same architecture are trained individually, and they show distinctive behaviours when applied to the same test data. Step‐by‐step analysis of each of them reveals that training the network with additional random noise injected into both the input and the output channel of the desired signal can lead to a decent prediction of the coherent noise based on good learning of the global features and, in the meantime, preserve almost all the data information from being lost. We propose this key lesson we learnt as a new method to improve the network's signal fidelity for shot‐domain seismic interference attenuation, which is essentially a signal separation task. Its effectiveness is demonstrated on field data from Africa with a comparison to a conventional physics‐based seismic interference attenuation method used in production. [ABSTRACT FROM AUTHOR]

Full‐waveform inversion, a popular technique that promises high‐resolution models, has helped in improving the salt definition in inverted velocity models. The success of the inversion relies heavily on having prior knowledge of the salt, and using advanced acquisition technology with long offsets and low frequencies. Salt bodies are often constructed by recursively picking the top and bottom of the salt from seismic images corresponding to tomography models, combined with flooding techniques. The process is time consuming and highly prone to error, especially in picking the bottom of the salt. Many studies suggest performing full‐waveform inversion with long offsets and low frequencies after constructing the salt bodies to correct the misinterpreted boundaries. Here, we focus on detecting the bottom of the salt automatically by utilizing deep learning tools. We specifically generate many random one‐dimensional models, containing or free of salt bodies, and calculate the corresponding shot gathers. We then apply full‐waveform inversion starting with salt flooded versions of those models, and the results of the full‐waveform inversion become inputs to the neural network, whereas the corresponding true one‐dimensional models are the output. The network is trained in a regression manner to detect the bottom of the salt and estimate the subsalt velocity. We analyse three scenarios in creating the training datasets and test their performance on the two‐dimensional BP 2004 salt model. We show that when the network succeeds in estimating the subsalt velocity, the requirement of low frequencies and long offsets are somewhat mitigated. In general, this work allows us to merge the top‐to‐bottom approach with full‐waveform inversion, save the bottom of the salt picking time and empower full‐waveform inversion to converge in the absence of low frequencies and long offsets in the data. [ABSTRACT FROM AUTHOR]

We propose a method to generate seismic images with corresponding fault labels for augmenting training data in automatic fault detection. Our method is based on two generative adversarial networks: one for creating a fault system and the other for generating two‐dimensional seismic images with faults as a condition. Our method can capture the characteristics of field seismic data during inference to generate samples that have properties of both field seismic data and synthetic training data. We then use the newly generated seismic images with corresponding fault labels to train a convolutional neural network for fault picking. We test the proposed approach on a three‐dimensional field dataset from the Gulf of Mexico. We use different areas in the field dataset as input to generate new training data for corresponding fault‐picking models. The results show that the generated training data from our method help in improving the fault‐picking models in the targeted areas. [ABSTRACT FROM AUTHOR]

The distribution of Oligo--Miocene magmatic rocks from southern Tibet in space and time yields critical information on the geometry and deformation of the subducted Indian lithosphere which impacts on plateau growth following the India and Eurasia collision. A growing body of geophysical evidence has shown that the subducted Indian lithosphere beneath the Tibetan Plateau has been torn apart. However, the spatiotemporal distribution and cause of the tearing remain enigmatic. Timing of the post-collisional magmatic rocks in southern Tibet exhibits four patterns of decreasing ages; magmatism began earlier in the west and east Himalayan syntaxis and evolved to two age undulations in the central southern Tibet. Seismic images show that regions of slab window (both 90°E and 84°E) and flattened subducted lithosphere (both 86°E and 81°E) are present at depth of 135 km. Correspondingly, increasing mineral crystallization temperatures (absolute value of 50 °C) were recorded in the Oligo--Miocene ultrapotassic-potassic rocks at 90°E and 84°E, while opposing trends were shown by coeval ultrapotassic-potassic rocks at 86°E and 81°E. Besides, the melting depth of the Oligo--Miocene ultrapotassic-potassic primitive melts decreases from nearly 100 km to 70 km between 81°E and 90°E, probably indicating progressive rising of the lithosphere-asthenosphere boundary. Such variations were possibly the results of the focused flow and upwelling of asthenosphere, which advanced rapidly but diachronously through weakened and torn sectors within the overlying Indian slab. The upwellings probably induced diachronously upward bending of the residual Indian slab and its flattening, which accelerated the tearing of the Indian lithosphere during continental subduction. [ABSTRACT FROM AUTHOR]

Widespread mud eruptions have been identified along the northern margin of the South China Sea. However, the ultimate driving force for initiation of the eruptions remains to be discussed. Here, we investigated the discovered hydrothermal vents and the presence of gas/fluid occurring above deepseated igneous intrusions and volcanic edifices and consider them as a likely trigger. Using a one-dimensional (1D) numerical model, SILLi 1.0, we studied the thermal effect of a 100-m-thick dolerite sill complex emplaced at 1100 °C and 3.8 km depth within the organic-rich Triassic series (~3.5 wt% total organic carbon). The modeled scenarios also account for variations in sill count and thickness and the effects of erosion. The calculated CH4 produced from contact metamorphism reaches a peak of ~700 kg/m²/yr shortly after emplacement of the first sill in the complex. When scaled to a sill size of 100 m by 25 km², i.e., a sill volume of 2500 km³, ~42-550 Mt of CH4, depending on individual sill thickness, have been produced within 10,000 yr of sill emplacement from sedimentary organic matter in the aureole, with peak generation of ~17.5 Mt CH4/yr just after emplacement. The newly discovered vents are suggested to have formed as a direct result of sill emplacement and by overpressure generated in the sediments around the sill by hydrothermal fluids and gases liberated during organic matter degradation. Combining data from seismic images with the likely production of carbon gas released from contact metamorphism with igneous intrusions, we suggest that volcanically induced gas release is common along the South China Sea margin. This study supports the important role of magmatism in driving regional-scale gas/fluid seepage. Our findings also suggest that the hydrothermal vents and intrusive magmatic bodies provide postemplacement fluid focusing pathways to shallower strata or the sea-floor over long time scales. [ABSTRACT FROM AUTHOR]

Data-driven reduced order models (ROMs) have recently emerged as an efficient tool for the solution of inverse scattering problems with applications to seismic and sonar imaging. One requirement of this approach is that it uses the full square multiple-input/multiple-output (MIMO) matrix-valued transfer function as the data for multidimensional problems. The synthetic aperture radar (SAR), however, is limited to the single-input/single-output (SISO) measurements corresponding to the diagonal of the matrix transfer function. Here we present a ROM-based Lippmann-Schwinger approach overcoming this drawback. The ROMs are constructed to match the data for each source-receiver pair separately, and these are used to construct internal solutions for the corresponding source using only the data-driven Gramian. Efficiency of the proposed approach is demonstrated on 2D and 2.5D (3D propagation and 2D reflectors) numerical examples. The new algorithm not only suppresses multiple echoes seen in the Born imaging but also takes advantage of their illumination of some back sides of the reflectors, improving the quality of their mapping. [ABSTRACT FROM AUTHOR]

Subducting the buoyant crustal material of an aseismic oceanic ridge has been regarded as a dominant contributor to flat slab subduction. However, normal‐dip subduction is also observed in some cases where ridges are subducting. In this study, we compare the subduction of two ridges on the Nazca Plate: Nazca Ridge (flat slab) and Iquique Ridge (normal‐dip slab). Anisotropy determined by shear wave splitting observation suggests that the low‐velocity anomalies found beneath the ridges are mapping anisotropic structure into isotropic velocities. After a tomographic inversion incorporating anisotropy models for both ridges, we find that the low‐velocity anomalies found beneath the Nazca Ridge are not anisotropic and therefore likely represent warm mantle, and those beneath the Iquique Ridge are caused by anisotropy. We conclude that subslab mantle buoyancy has a larger impact on the subduction angle than the crustal material of the ridge. Plain Language Summary: Understanding the subduction process and how the mantle flows is pivotal in understanding the planetary evolution of Earth. Yet, several subduction characteristics remain unsolved, and the angle of plate subduction, which is often categorized into <30° (shallow), ∼30–35° (normal), >35° (steep), is one of those. Subduction of an oceanic ridge on a plate has been proposed as a cause of a shallow subduction angle since the thick crust of the ridge is less dense than the surrounding mantle. However, normal angles have been observed in some cases where oceanic ridges are subducting. In this study, we compare the Nazca Ridge and the Iquique Ridge, on the Nazca Plate subducting beneath South America. The subduction angles of the Nazca and Iquique Ridges are shallow and normal, respectively. When we incorporate directional variations in seismic wave velocities, which are produced by mantle flow, in seismic tomographic imaging, we find that the subducting oceanic ridge may not be a primary factor producing shallow angle subduction. Instead, the warm mantle beneath the Nazca Ridge may provide the buoyancy to support the Nazca Plate. Comparably, since the mantle beneath the Iquique Ridge is not warm, the subduction angle would stay normal. Key Points: Despite having different subduction angles, subslab low‐velocity anomalies are found beneath both Nazca Ridge and Iquique RidgeWe introduced hypothetical seismic anisotropy in tomographic inversion to explore the origin of the subslab low‐velocity anomaliesThe low‐velocity anomalies beneath the Nazca Ridge and Iquique Ridge may come from buoyant warm mantle and anisotropy, respectively [ABSTRACT FROM AUTHOR]

We conducted a time‐lapse seismic experiment utilizing automated active seismic source and sensor arrays to monitor a reactivated fault within the Opalinus clay formation at the Mont Terri Rock Laboratory (Switzerland), an analog caprock for geologic carbon storage. A series of six brine injections were conducted into the so‐called Main Fault to reactivate it. Seismic instrumentation in five monitoring boreholes on either side of the fault was used to continuously probe changes in P‐wave travel‐times associated with fault displacement and leakage. We performed time‐lapse travel‐time tomography on five hundred sequential data sets; this revealed a zone of decreased P‐wave velocity, up to 16 m/s, during each injection cycle, followed by a velocity increase during shut‐in. These results demonstrate varying elastic property perturbations, both spatially and temporally, along the fault plane during reactivation. We then interpreted these velocity changes in terms of fault dilation induced by pressurized fluids along the fault. Plain Language Summary: Faults within clay formation caprocks for CO2 storage reservoirs are possible pathways for leakage and loss of containment. Understanding how these faults in clay‐rich rocks reactivate and leak fluids is important for predicting, detecting, and preventing CO2 movement. Passive seismic monitoring is challenging because of the lack of observable seismic events in such clay‐rich fault rupture. In this study, we measure changes in P‐wave velocity to monitor a fault reactivated by brine injections directly into the Main Fault at the Mont Terri Rock Laboratory, Switzerland. We use a recently developed time‐lapse seismic technique called Continuous Active‐Source Seismic Monitoring (CASSM), which allows us to make these measurements within a few minutes and observe small changes on the same timescale. We relate the measured changes in P‐wave velocity to the opening of that fault damage zone by using a rock physics model, which helps explain changes in permeability within the fault zone. Key Points: measuring p‐wave velocity changes during fault reactivationmonitoring fault reactivation in an analog caprock for geologic carbon storagefracture damage zone modeling from p‐wave velocities [ABSTRACT FROM AUTHOR]

[Display omitted] After two continents collide, plate convergence and orogenesis are sustained because subducted continental lithosphere continues pulling the surface plate. It remains controversial how, why, and when continental plate convergence and collision slow down and eventually cease. We use an unprecedented data coverage and present a regional-scale seismic tomographic image of the mantle structure beneath the Tibetan Plateau. In the mantle transition zone, we identify multiple high-velocity anomalies and interpret them as detached pieces of the Indian continental slab. Facilitated by internal heterogeneity of the continental lithosphere, piecewise slab detachments could reduce the slab pull force, resulting in the Miocene slowdown of the India-Eurasia convergence and coeval diachronous potassic volcanism in southern Tibet. We propose that slab detachment is a mechanism that eventually will lead to the end of the Indo-Eurasian continental collision and the Himalayan orogeny. [ABSTRACT FROM AUTHOR]

Much of our knowledge of the North American lithosphere comes from imaging seismic velocities. Additional constraints on the subsurface can be gained by studying seismic attenuation, which has different sensitivity to physical properties. We produce a model of lateral variations in attenuation across the conterminous U.S. by analyzing data recorded by the EarthScope Transportable Array. We divide the study area into 12 overlapping tiles and differential attenuation is measured in each tile independently; and twice for four of the tiles. Measurements are combined into a smooth map using a set of linear inversions. Comparing results for adjacent tiles and for repeated tiles shows that the imaged features are robust. The final map shows generally higher attenuation west of the Rocky Mountain Front than east of it, with significant small length scale variations superimposed on that broad pattern. In general, there is a strong anticorrelation between differential attenuation and shear wave velocities at depths of 80–250 km. However, a given change in velocity may correspond to a large or small change in attenuation, depending on the area; suggesting that different physical mechanisms are operating. In the western and south‐central U.S., as well as the Appalachians, velocity variations are large compared to attenuation changes, while the opposite is true in the north‐central and southeastern U.S. Calculations with the Very Broadband Rheology calculator show that these results are consistent with the main source of heterogeneity being temperature and melt fraction in the former regions and grain size variability in the latter ones. Plain Language Summary: Seismic waves in the mantle propagate at lower speeds when temperatures are higher, rocks have higher water content, and small amounts of melt are present. These conditions also affect how much energy the wave loses as it passes through, which we call seismic attenuation. In this study we produce a map of seismic attenuation for the conterminous United States. We find that, in most places, where seismic velocities are low, attenuation is high, and vice versa. This is what we expect. Interestingly, the size of the change in attenuation that corresponds to a given change in velocity varies by region. In places with thicker lithosphere and without recent tectonic activity attenuation anomalies are large compared to velocity anomalies, the opposite is true in places with thin lithosphere and recent tectonic activity. Considering the results of lab experiments on velocity and attenuation, this suggests that in the regions with thick lithosphere and without recent tectonic activity the main cause of the anomalies is changes in the size of the mineral grains in the mantle, whereas in the regions with thin lithosphere and recent tectonic activity the main cause of the anomalies is changes in temperature and the amount of melt. Key Points: Teleseismic P attenuation is largely anti‐correlated with upper mantle seismic velocitySignificant lateral variations in attenuation are observed in the tectonically quiescent regionsThe ratio of attenuation to velocity anomaly amplitudes may be indicative of mantle conditions [ABSTRACT FROM AUTHOR]

Q-compensated reverse time migration (Q-RTM) is a crucial technique in seismic imaging. However, stability is a prominent concern due to the exponential increase in high-frequency ambient noise during seismic wavefield propagation. The two primary strategies for mitigating instability in Q-RTM are regularization and low-pass filtering. Q-RTM instability can be addressed through regularization. However, determining the appropriate regularization parameters is often an experimental process, leading to challenges in accurately recovering the wavefield. Another approach to control instability is low-pass filtering. Nevertheless, selecting the cutoff frequency for different Q values is a complex task. In situations with low signal-to-noise ratios (SNRs) in seismic data, using low-pass filtering can make Q-RTM highly unstable. The need for a small cutoff frequency for stability can result in a significant loss of high-frequency signals. In this study, we propose a multi-task learning (MTL) framework that leverages data-driven concepts to address the issue of amplitude attenuation in seismic records, particularly when dealing with instability during the Q-RTM (reverse time migration with Q-attenuation) process. Our innovative framework is executed using a convolutional neural network. This network has the capability to both predict and compensate for the missing high-frequency components caused by Q-effects while simultaneously reconstructing the low-frequency information present in seismograms. This approach helps mitigate overwhelming instability phenomena and enhances the overall generalization capacity of the model. Numerical examples demonstrate that our Q-RTM results closely align with the reference images, indicating the effectiveness of our proposed MTL frequency-extension method. This method effectively compensates for the attenuation of high-frequency signals and mitigates the instability issues associated with the traditional Q-RTM process. [ABSTRACT FROM AUTHOR]

Active or man-made seismic sources are often used to delineate subsurface geological structures via seismic imaging techniques. Although conventional controlled seismic sources have provided high-resolution subsurface images, the high cost of data acquisition necessitates improved use of cheaper alternative seismic sources for subsurface imaging, such as traffic noise, quarry blasts and earthquakes. Here, we explore the potential use of quarry blasts as an active seismic source for subsurface imaging. The goal of our study is to demonstrate the utility of quarry blasts as seismic sources for investigating subsurface structure with a field experiment in Mill Creek, southern Oklahoma. We deployed a 7455-m-long receiver line of 72 receivers (105 m spacing) extending away from a quarry, with a recording time of 28 d. We analyse similarities between blasts, wave types, origin time and excitation location of quarry blasts, which is important information for using such signal as an active source. Given a relative homogeneous geological condition in the study area, we could not observe clear reflected waves. However, the use of P - and S -wave velocities led to the delineation of a major basement damage zone in the study area, which further demonstrates the viability of quarry blasts as a useful seismic source for subsurface geological imaging. [ABSTRACT FROM AUTHOR]

We investigate the role of crustal architectures playing in controlling the genesis and nature of supergiant Jiaodong orogenic gold system via seismic imaging. We deployed an NWW-SEE-trending broadband seismic linear array traversing the province to construct a shear wave velocity (VS) transect based on ambient noise tomography. Our crustal VS transect reveals a prominent low velocity zone (LVZ) at 12–20 km depths across the whole province and high VS anomalies at 8–12 km depths underneath the western part that hosts most of the ore tonnage. We interpret the former as a hydrothermal alteration zone related to mineralization and the latter as a crustal component containing large amounts of mafic rocks (e.g., amphibolites). In addition, a listric fault system and a strike-slip fault system are imaged in the western and eastern parts of the province, respectively. Combining features of crustal velocities with regional geological observations that ores are tempo-spatially associated with mafic dikes, we propose a model of gold metallogenic processes. In this model, ponding and degassing of the mafic magmas led to the formation of a hydrothermal alteration LVZ in the middle crust prior to mineralization. Later, auriferous fluid was released from this intra-crustal hydrothermal alteration zone as it was heated by upwelling asthenosphere. The ore fluid ascending along the listric fault system possibly leached extra gold from the upper-crustal amphibolites. Then, the gold-rich fluid migrated up along different fault systems, leading to the disseminated-veinlet mineralization with great gold endowment in the western listric fault system and the quartz-vein type mineralization with relatively small ore tonnage in the eastern strike-slip fault system. [ABSTRACT FROM AUTHOR]

High-quality seismic data is key to determining the success or failure of oil and gas exploration in complex mountain fronts. Observation systems are important factors affecting the quality of the seismic collection. Currently, the seismic collection design of piedmont zones is mainly based on optimizing and analyzing a single collection parameter. The multiparameter optimization is not performed as a whole. Moreover, the collection parameters that significantly impact the imaging quality of piedmont zones are not deeply excavated. This paper first uses the relationship between the acquisition parameters and the signal-to-noise ratio of a previous seismic data analysis to solve this problem and then obtains the prestack offset profiles of different acquisition parameters through forward modeling and prestack offset methods, further analyzes the acquisition parameters that have a more significant impact on imaging quality, uses the degenerative treatment method to process and analyze the newly collected seismic data, and delves into the most sensitive acquisition parameters for imaging quality. The results show that the influence of the gun line distance is the largest, and the influence of the receiving line distance, the number of receiving lines, and the number of receiving channels is reduced. Therefore, the effect of each acquisition parameter on the imaging quality is different, and different acquisition parameters should be selected according to the depth of the target layer. [ABSTRACT FROM AUTHOR]

The salt body is one of the important structures in reservoirs, but there are still challenges in fully interpreting salt bodies from three-dimensional seismic data. In actual exploration and production operations, the characteristics of salt bodies in most three-dimensional seismic slices are not obvious, making it difficult to obtain labeled samples for salt bodies. The complex structural features of salt bodies, which are similar to background features, severely interfere with the accuracy of salt body interpretation, leading to prominent issues of multi-solution prediction. To address these problems, we treat salt body interpretation as a semantic segmentation problem of three-dimensional seismic images and propose 3D Salt-Net. We have designed a self-training paradigm based on optical flow estimation and multi-level constraints to overcome the difficulty of obtaining labeled samples for salt bodies and provide high-quality pseudo-labels. Based on this, we establish a mapping relationship between three-dimensional seismic data and three-dimensional salt body masks to ensure the spatial continuity and smoothness of salt body predictions. In addition, we have designed two plug-and-play salt body attention modules to collectively address the issues of blurred edge features of salt bodies and interference from background similarity in the prediction results. We conducted experiments on the SEAM and F3 seismic datasets, using only 3% of the labels for training, while the rest were used for validation. Experimental results demonstrate that the 3D Salt-Net method outperforms previous state-of-the-art methods in terms of salt body segmentation and achieves satisfactory results. [ABSTRACT FROM AUTHOR]

Seismic investigations were performed at a site in the southwest of Sweden where major quick‐clay landslides have occurred in the past. Given the potential high risk of the area and the presence of medium infrastructures, the site posed a need for detailed investigations in a wide depth range and in high resolution. A high‐fold seismic survey was designed and conducted along two profiles using a 1–2 m receiver and shot spacing in order to retrieve both P‐ and S‐wavefield seismic images from vertical component data. The data were analysed by combining first‐break traveltime tomography and surface‐wave analysis as well as P‐ and S‐wavefield reflection seismic imaging. Using the first breaks, P‐wave velocity (VP) models were estimated, indicating the bedrock topography along the profiles and the sediment characteristics. The S‐wave velocity (Vs) models were estimated from the surface waves and indicated areas of low shear strength. Combined with VP and Vs models, this permits the estimation of VP/VS, a parameter that can indicate areas with high water content, significant for the detection of quick clays and possible liquefaction issues. The results are integrated with the P‐ and S‐wave reflection seismic images and compared with other geophysical investigations, such as magnetic and gravity data that were collected along the profiles. [ABSTRACT FROM AUTHOR]

The Revell Site, located in Northwestern Ontario within the Canadian Shield, is being assessed as a potential Deep Geological Repository (DGR) for Canada's used nuclear fuel. Effective DGR establishment requires comprehensive subsurface assessment, particularly in evaluating structural stability, hydrogeological attributes, geological composition, and geochemical properties. Key among these considerations is understanding the three-dimensional characteristics of structural features to ensure the site's suitability for long-term containment of radioactive materials. This case study focuses on imaging and characterizing structures within the predominantly intact biotite granodiorite-tonalite host rock at the Revell Site. Borehole data reveals these structures as mostly sub-horizontal mafic intrusions, with thicknesses of up to 3.5 m. They appear either as separate, discrete entities or in stacked configurations. Despite their limited thickness, most of these intrusions exhibit discernible attributes in surface seismic images due to their distinct physical properties. Some, however, exhibit velocity variations that decrease their overall reflectivity properties. An integrative approach employing surface seismic, Vertical Seismic Profiles (VSP), and borehole data facilitates the spatial identification of over 30 of these mafic structures. This comprehensive characterization not only lays the foundational framework for future discrete fracture network models but also provides important support for simulations related to fluid flow dynamics, groundwater behavior, contaminant dispersion, and heat transport mechanisms within the Revell Site. This study underscores the critical role of in-depth subsurface characterization in ensuring the secure, long-term management of radioactive materials in DGRs. [ABSTRACT FROM AUTHOR]

There is an ongoing debate on the processes producing background seismicity and deformation transients across seismic gaps, i.e., regions that lack historical large-magnitude earthquakes. Essential missing elements are geophysical images that resolve sources of geophysical unrest. Here, we apply seismic scattering and absorption tomography to data recorded during the 2010–2014 seismic sequence within the Mt. Pollino seismic gap region (Southern Italy). The tomographic models show high sensitivity to fluid content, deformed fractured structures, and impermeable layers stopping fluid migrations. They bridge the gaps between geological and geophysical models and provide a highly-resolved image of the source of seismic and deformation unrest within this seismic gap. High absorption topping the western Pollino seismic volume appears pressurized between the low-Vp/Vs and low-scattering San Donato metamorphic core and a deep basement. Absorbing fluids can only migrate laterally to the east, blocked in the west and southwest by deep low-scattering barriers associated with east-dipping faults and to the north and southeast by saturated overpressurized low-scattering basins. This eastern migration is only partially effective, producing seismicity across the lowest boundary of the high-absorption volume. Our results showcase the potential of seismic scattering and absorption when imaging structures causing geophysical unrest processes across fault networks. [ABSTRACT FROM AUTHOR]

The Southern San Andreas Fault (SSAF) in California is one of the most thoroughly studied faults in the world, but its configuration at seismogenic depths remains enigmatic in the Coachella Valley. We use a combination of space geodetic and seismic observations to demonstrate that the relatively straight southernmost section of the SSAF, between Thousand Palms and Bombay Beach, is dipping to the northeast at 60–80° throughout the upper crust (<10 km), including the shallow aseismic layer. We constrain the fault attitude in the top 2–3 km using inversions of surface displacements associated with shallow creep, and seismic data from a dense nodal array crossing the fault trace near Thousand Palms. The data inversions show that the shallow dipping structure connects with clusters of seismicity at depth, indicating a continuous throughgoing fault surface. The dipping fault geometry has important implications for the long‐term fault slip rate, the intensity of ground shaking during future large earthquakes, and the effective strength of the southern SAF. Plain Language Summary: The San Andreas Fault (SAF) is capable of large, destructive earthquakes and poses significant seismic hazard in California. In the Coachella Valley, the geometry of the SAF beneath the Earth's surface has been the subject of much debate. Here, we present new constraints on the fault's structure in the uppermost 2–3 km of the crust from several new sets of observations. We measure slow movement of the fault at the surface using satellite‐based radar and perform simulations of what fault geometry is most likely to produce the observed motion. We also installed an array of sensors across the SAF to measure seismic waves that travel through or reflect off the fault. Analysis of both data sets indicates that the shallow portion of the SAF dips to the northeast in the Coachella Valley. Precise locations of small earthquakes at greater depths exhibit a similar trend—we thus suggest the fault dips throughout the Earth's crust. Better understanding the geometry of this section of the SAF has implications for earthquake hazards in Southern California, as well as the fault's geologic history. Key Points: InSAR measurements and models of shallow creep on the southernmost San Andreas Fault indicate a moderate‐to‐steep northeast fault dipSeismic imaging using data from a dense array deployed near Thousand Palms reveal northeast dipping fault damage zonesA joint interpretation of available data suggests that the northeast dip persists throughout the upper 10 km of the crust [ABSTRACT FROM AUTHOR]

We provide a comprehensive image of the mantle transition zone (MTZ) beneath northeastern China by performing Variable Bin Radius Stacking of receiver functions. A massive seismic data set consisting of over 133 000 receiver functions recorded by 1208 broad-band stations is processed. Our results reveal fine-scale topography on the 410- and 660-km discontinuities defining the upper and lower bounds of the MTZ, lateral variations in the MTZ thickness and slab interfaces within the MTZ. In particular, unambiguous images of the slab interfaces provide direct evidence for the presence of the stagnant Pacific subducting slab below the eastern portion of the study area. A widespread deepening of the 410-km discontinuity is consistent with a hot and wet low-velocity upper mantle resulting from dehydration of the stagnant slab. Prominent depressions are evident in the depth to the 660-km discontinuity, with a thickened MTZ associated with the cold stagnating slab. Localized uplifts of the 660-km discontinuity are possibly caused by partial melt under the slab. These features attest to the influence of the Pacific plate on the MTZ. Additionally, a pronounced upwarp on the 660-km interface with a thin MTZ agrees with a previously hypothesized mantle upwelling through a slab window, possibly triggered by the sinking of the stagnant slab. Moreover, the western part of the study region is characterized by alternating ups and downs of the 410-km interface, while the topography of the 660-km discontinuity is relatively flat. We propose the western region is dominated by foundering of delaminated lithospheric blocks that induced upward mantle return flows upon entrance into the MTZ. [ABSTRACT FROM AUTHOR]

Many seismic imaging methods use wavefield extrapolation operators to redatum sources and receivers from the surface into the subsurface. We discuss wavefield extrapolation operators that account for internal multiple reflections, in particular propagator matrices, transfer matrices and Marchenko focusing functions. A propagator matrix is a square matrix that 'propagates' a wavefield vector from one depth level to another. It accounts for primaries and multiples and holds for propagating and evanescent waves. A Marchenko focusing function is a wavefield that focuses at a designated point in space at zero time. Marchenko focusing functions are useful for retrieving the wavefield inside a heterogeneous medium from the reflection response at its surface. By expressing these focusing functions in terms of the propagator matrix, the usual approximations (such as ignoring evanescent waves) are avoided. While a propagator matrix acts on the full wavefield vector, a transfer matrix (according to the definition used in this paper) 'transfers' a decomposed wavefield vector (containing downgoing and upgoing waves) from one depth level to another. It can be expressed in terms of decomposed Marchenko focusing functions. We present propagator matrices, transfer matrices and Marchenko focusing functions in a consistent way and discuss their mutual relations. In the main text we consider the acoustic situation and in the appendices we discuss other wave phenomena. Understanding these mutual connections may lead to new developments of Marchenko theory and its applications in wavefield focusing, Green's function retrieval and imaging. [ABSTRACT FROM AUTHOR]

We use a newly developed 2-D elastic reverse time migration (RTM) imaging algorithm based on the Helmholtz decomposition to test approaches for imaging the descending slab in subduction zone regions using local earthquake sources. Our elastic RTM method is designed to reconstruct incident and scattered wavefields at depth, isolate constituent P- and S- wave components via Helmholtz decomposition, and evaluate normalized imaging functions that leverage dominant P and S signals. This method allows us to target particular converted-wave scattering geometries, for example incident S to scattered P , which may be expected to have dominant signals in any given data set. The method is intended to be applied to dense seismic array observations that adequately capture both incident and converted wavefields. We draw a direct connection between our imaging functions and the first-order contrasts in shear wave material properties across seismic discontinuities. Through tests on synthetic data using either S → P or P → S conversions, we find that our technique can successfully recover the structure of a subducting slab using data from a dense wide-angle array of surface stations. We also calculate images with a small-aperture array to test the impact of array geometry on image resolution and interpretability. Our results show that our imaging technique is capable of imaging multiple seismic discontinuities at depth, even with a small number of earthquakes, but that limitations arise when a small aperture array is considered. In this case, the presence of artefacts makes it more difficult to determine the location of seismic discontinuities. [ABSTRACT FROM AUTHOR]

We present compact reverse time migration (CRTM), a real-time ultrasound imaging method that can exploit the full waveform information of ultrasonic wave records for imaging breast tissue. Conventional reverse time migration (RTM) computes the gradient of the reflective ultrasound data with respect to the perturbation of the velocity model of the soft tissues and the gradient can indicate the interface between different types of body tissue. In contrast to conventional reflection ultrasound (B-mode), which is based on the high-frequency approximation to the wave equation, the RTM algorithm is based on the complete wave equation, and can thus exploit the full waveform (wide-spectrum) information of the data and provide an image with higher resolution. Unfortunately, the computational burden of RTM is noticeably higher than the ray-based B-mode. This precludes real-time applications, one of the most important features of ultrasound imaging. The proposed CRTM algorithm can significantly reduce the computational costs of RTM, such that it can be applied for real-time imaging. We demonstrate the performance of CRTM through a synthetic experiment of ultrasound breast imaging. CRTM can be potentially adapted to related signal-processing fields, such as seismic imaging, acoustic camera systems, and radar imaging. [ABSTRACT FROM AUTHOR]

Alaska is located at the northernmost point of the interface between the Pacific plate and the North American continent. The subduction of the Pacific plate generates arc volcanoes along the Aleutian trench, which stops to the east at the Denali Volcanic Gap. This volcanic gap has been linked to the underthrusting of the Yakutat terrane, which might alter the thermal state of the mantle wedge and prevent melt formation. This implies that the limits of the volcanic activity should mirror the extent of the Yakutat subduction. However, the transition from the Pacific slab to the Yakutat terrane at depth is not fully understood. To investigate this issue, we processed a new composite seismic data set from six arrays deployed in the region from 2000 to 2018. We apply a multi-mode 3D Kirchhoff migration to obtain high-resolution 3D scattering images of the region. Our results highlight a sharp lateral boundary in the slab structure, with a 10 km Moho step, just offshore Anchorage, and a more gradual slab transition beneath the southern part of the Kenai peninsula. Our images from the Yakutat slab plunge down to 150 km depth are consistent with previous estimates of the Yakutat slab extent below the Alaska Range. Although the steeply dipping boundaries of the subducting Pacific lithospheres are not fully recovered, deep coherent signals from the Pacific slab are observed down to 150 km depth. These observations suggest that the crust is still partially uneclogitized at these depths in both slabs. [ABSTRACT FROM AUTHOR]

We consider an $ \ell $1-regularized problem, which has many applications such as compressed sensing, MRI, and seismic images. Although there have been some methods for the problem, they cannot obtain a highly accurate solution. To overcome this drawback, we propose a dual active set method, which is an active set method for the Fenchel dual problem of the $ \ell $1-regularized problem. Our method finds a highly accurate solution quickly because of the following features of the dual problem: (i) the dimension of the decision variables is smaller; (ii) since the optimal solution of the primal problem is sparse, the active constraints of the optimal solution of the dual problem are few and can be identified quickly. Then, we show that the method outputs the solution with finite iterations when it solves certain subproblems exactly. Moreover, we give its efficient and concrete implementation for Basis Pursuit Denoising (BPDN), which is a typical $ \ell $1-regularized problem. In numerical experiments for BPDN, our method provides more accurate solutions than all baselines in comparable computing time. [ABSTRACT FROM AUTHOR]

A new technique for P‐wave receiver function (PRF) inversion, within the framework of wave equation‐based adjoint tomography and referred to as receiver function adjoint tomography (RFAT), has been developed to obtain models of Vp, Vs, and density. This innovative technique fits the synthetic PRFs with observed PRFs and can better image the lateral variations of Vs from the crust to the uppermost mantle than traditional 1‐D PRF inversion. We utilized RFAT to perform high‐resolution imaging beneath a dense seismic array in Southeastern Tibet, revealing low‐velocity zones extending from the uppermost mantle to the crust, as well as an eastward dipping Moho under the Red River Fault (RRF). Our inversion results provide direct evidence for the existence of a distinct asthenospheric upwelling channel beneath the RRF, and further highlight the effectiveness of RFAT for accurately imaging subsurface structures. Plain Language Summary: We have developed a novel technique that allows for direct inversion of P‐wave receiver functions for a detailed 3‐D S‐wave velocity model. By applying this technique to a dense seismic array located in the southeastern margin of the Tibetan Plateau, we can resolve detailed crustal and uppermost mantle structures with strong lateral variations. The results indicate that the boundary faults of the geological blocks play a critical role in controlling the lateral variations of the crustal and upper mantle structures. Our findings provide new insights into the dynamics of lithosphere‐asthenosphere interactions and continental collision processes in this region. Key Points: A new method of 3‐D adjoint tomography for receiver function inversion is developedLateral heterogeneities of crust and uppermost mantle can be well resolved by this methodThe method reveals vertically extending low‐velocity zones under the Red River Fault due to mantle upwelling [ABSTRACT FROM AUTHOR]

The Yinggehai Basin (YB) in the northwest of the South China Sea (SCS) has preserved the complete evolution of the Red River Fault (RRF), whose motion over time has largely contributed to shaping the current tectonic framework of the southeastern Tibetan Plateau and Indochina Block. Here we, for the first time, image the RRF, crustal architecture, and crust‐mantle boundary (i.e., Moho) in detail under the YB from two high‐resolution 40‐km‐depth seismic reflection pre‐stack migration profiles. Our seismic reflection images provide direct seismic evidence to support a deep‐penetrating RRF, most likely crosscutting the Moho. In this sense, the RRF was a large‐scale shear zone, along with the clockwise rotation of the Indochina Block, to accommodate the Indochina Block extrusion during Oligocene‐early Miocene and form the YB. We highlight that the material extrusion along deep‐penetrating faults dominated the Tibetan Plateau deformation in response to the India‐Asia collision at an early stage. Plain Language Summary: The India‐Asia collision has not only created the highest Tibetan Plateau but also resulted in material extrusion outward along multiple large‐scale strike‐slip faults. The Red River Fault (RRF) is suggested to be one of these large‐scale strike‐slip faults. The geometry and deep extension of these strike‐slip faults are fundamental to understanding the manner of the material extrusion. Despite many studies describing geophysical and geological observations of the RRF, its nature in response to the India‐Asia collision remains unknown. The Yinggehai Basin (YB) northwest of the South China Sea with >10 km thick sedimentary has preserved complete evolution of the RRF. In this study, we use high‐resolution deep seismic reflection data with a depth of 40 km, for the first time, to map the crustal architecture, detailed geometries of the crust‐mantle boundary (i.e., Moho) and the RRF under the YB. The seismic reflection images indicate that the RRF is a deep‐penetrating fault that could serve as a large‐scale shear zone. We further highlight that the RRF played a vital role in moving the Indochina Block away from the Tibetan Plateau in response to the India‐Asia collision at an early stage, and subsequently stretching and thinning the continent, eventually forming the YB. Key Points: For the first time, we image a reflected Moho cut by the Red River Fault (RRF) structure under the Yinggehai Basin (YB), South China SeaThe RRF in the YB is inferred to be a deep‐penetrating fault, most likely cutting through the Moho boundaryThe RRF was inferred to be a large‐scale shear zone accommodating the Indochina Block extrusion during Oligocene‐early Miocene [ABSTRACT FROM AUTHOR]

Fault detection is an important step in subsurface interpretation and reservoir characterization from 3D seismic images. Due to the numerous and complex fault structures in seismic images, manual seismic interpretation is time-consuming and requires intensive work. We applied a pre-trained CNN model to predict faults from the 3D seismic volume of the Poseidon field in the Browse Basin, Australia. This field is highly structured with complex normal faulting throughout the targeted Plover Formations. Our motivation for this work is to compare machine-learning-based fault prediction to user-interpreted fault identification supported by seismic variance attributes. We found reasonably satisfactory results using CNN with an improved fault probability volume that outperforms variance technology. Therefore, we propose that this workflow could reduce time and be able to predict faults quite accurately in most structurally complex areas. [ABSTRACT FROM AUTHOR]

The internal structure and architecture of evaporite sequences is often overlooked, with attention frequently concentrating on the external geometries that salt bodies form. The availability of extensive 3D seismic data affords the opportunity to interpret the internal structures within these evaporite sequences and comprehensively characterise the different structural facies over large areas. This paper concentrates on the Zechstein Supergroup evaporite deposits within the Southern North Sea of the United Kingdom's Continental Shelf. This analysis of the internal structural complexity and stratigraphic heterogeneity utilises 26,000 km2 of 3D seismic data together with over 96 wells from the Southern North Sea. Characterisation of the different structural facies present was undertaken alongside mapping their spatial distribution to understand the relationship they have with one another and the structural evolution that may have been taken. This work has (1) characterised and mapped six different internal structural facies present within the Zechstein with increasing levels of deformation; (2) shown the internal lithological heterogeneity is indicative of variations in the vertical strength profile of layered evaporite sequences; (3) discontinuous high‐amplitude reflections within the Zechstein are as a result of the geometries being too steeply dipping for the seismic data to image; and (4) the ability to possibly predict the internal heterogeneity of areas of poorly imaged salt, such as within large diapiric salt structures, from surrounding structural facies. These findings suggest that there is significant internal complexity even within areas of the basin with minor mobilisation to the external salt geometry. [ABSTRACT FROM AUTHOR]

Simulating the temporal evolution of wavefield solutions through models with heterogeneous material properties is of practical interest for many scientific applications. The acoustic wave equation (AWE) is often used for studying wave propagation in both fluids and solids and is crucial for many applications including seismic imaging and inversion and non-destructive testing. Because analytical AWE solutions rarely exist for complex heterogeneous media, methods for generating numerical AWE solutions are very desirable. Traditional numerical solvers require discrete model representations with many restrictions placed on the shape and spacing of grid elements. This work uses a relatively new class of numerical solvers known as physics-informed neural networks (PINNs) that provide a mesh-free alternative for generating AWE solutions using a deep neural-network framework. We encapsulate a time-domain AWE formulation within a loss function that is used to train network parameters. The initial conditions are implemented by enforcing hard constraints on the neural network instead of including them as separate loss-function terms. We also use a Fourier neural network (FNN) to alleviate the spectral bias commonly observed when using fully connected neural network in the conventional PINN approach. Numerical tests on both 2D homogeneous and heterogeneous velocity models confirm the accuracy of our approach. We observe that using FNNs helps in the convergence of AWE solutions especially for heterogeneous models. We compare PINN-based solutions with those computed by the highly accurate conventional pseudo-spectral method, and observe that the normalized energy differences between the two sets of solutions were less than 4% for all numerical tests. [ABSTRACT FROM AUTHOR]

The Hawaiian Islands represent the youngest portion of the long‐lived Hawaiian‐Emperor Seamount chain sourced from hotspot volcanism on the northwesterly moving Pacific plate. Despite being one of the best studied hotspots on Earth, our understanding of the lithospheric discontinuity structure of Hawaiʻi is inhibited by a lack of a long‐term distributed broadband seismic network and lithospheric‐scale imaging of other oceanic hotspots. In this study, we analyze 3,530 teleseismic waveforms from 1,130 events recorded at 47 stations to compute P‐wave radial receiver functions. We then incorporate an elevation correction to adaptive common conversion point stacking to create a 3D discontinuity volume of the lithospheric structure associated with an ocean island hotspot. We find a positive amplitude conversion consistently at ∼12 km below sea level, likely representing the crust‐mantle transition below much of the island. However, this conversion occurs at only ∼6 km below Kīlauea, where local tomography images a zone of high P‐wave velocities (>7 km/s) within the crust and deformation of olivine phenocrysts within a magmatic reservoir is thought to occur. At this location, seismicity also extends vertically to 35 km depth before merging with an interpreted sill complex at ∼40 km. We interpret this anomalously shallow high amplitude discontinuity as an olivine‐dominated magmatic reservoir with an estimated maximum 5% melt over the resolution of our data at crustal depths. This represents the shallow seismic signature of a much more extensive magma system extending to depth. These images provide a better understanding of the magmatic structure associated with oceanic hotspot volcanoes. Plain Language Summary: The Hawaiian Islands are the product of hotspot volcanism on the northwesterly moving Pacific plate. Our understanding on the formation of Hawaiʻi and other oceanic hotspot volcanoes is reliant on our ability to discern the process of magma formation and transportation through the mantle and up to the volcanic caldera. We have a thorough understanding of broad mantle characteristics and crustal structure beneath Hawaiʻi but lack any relatively higher resolved lithospheric seismic image given the instrumentation and data that are available. In this study, we derive a 3‐D discontinuity volume beneath Hawaiʻi. We find a discontinuity consistently at ∼12 km below sea level which we interpret to be the Moho. Beneath Kīlauea, the most active caldera on Hawaiʻi, this discontinuity shallows to ∼6 km and is unlikely to represent the Moho. These depths are consistent with high (mantle‐like) seismic velocities found in previous seismic studies. Seismicity surrounds this shallow discontinuity and extends down to ∼35 km depth possibly connecting to another magma reservoir. We interpret the shallow discontinuity beneath Kīlauea as ultramafic material formed within the crust by extensive differentiation in a low‐melt magma reservoir. These images provide a better understanding of magmatic structure associated with oceanic hotspot volcanoes. Key Points: Elevation‐corrected P‐to‐S receiver functions are used to create a 3‐D discontinuity model beneath the main Hawaiian islandA high‐amplitude positive arrival is imaged ∼6 km BSL at Kīlauea spatially coinciding with seismicity extending vertically to ∼35 km depthThis shallow arrival represents the uppermost portion of an olivine‐dominated magmatic reservoir with up to ∼5% melt [ABSTRACT FROM AUTHOR]

Coastal protection in the form of dike constructions has a long history at the German North Frisian coast dating back to the High Middle Ages. As the vast majority of the dikes built prior to the devastating storm surges of the Middle Ages have been irretrievably destroyed, mostly sparse remains and only a few well preserved of these medieval dikes are found along the German North Frisian coast and within the Wadden Sea. Not all details of their construction and dimensions are yet understood. In the present case study, we investigate the historical Schardeich on the island of Pellworm in the German North Sea in a noninvasive way using shear waves (SH‐waves). For the data interpretation, we applied a combination of seismic full waveform inversion and classical seismic reflection imaging to determine the interior structure of the dike and its underlying layers at the highest possible resolution. The results obtained on land are compared with dike remains found in the tidal flats. These remains show up in marine seismic sections as characteristic reflections, which probably represent a compaction layer caused by the load of the former dike. For ground truthing, we compare the seismic results with internal dike structures found in nearby excavations. The comparison highlights that FWI is a reliable tool for near‐surface archaeological prospecting. We find that SH‐wave FWI provides decimetre‐scale velocity and density models that allow, together with the seismic reflection section, to determine distinct construction phases of the dike. The investigated dike further shows a depression at base level of about 0.75 m, which is of the same order as observed for the dike base reflections in the tidal flats. Transferring these findings to the dike remains mapped in the tidal flats, we derive a height of the former dike from 2.2 to 4.4 m. [ABSTRACT FROM AUTHOR]

Full-waveform inversion (FWI) is the current standard method to determine final and detailed model parameters to be used in the seismic imaging process. However, FWI is an ill-posed problem that easily achieves a local minimum, leading the model solution in the wrong direction. Recently, some works proposed integrating FWI with Convolutional Neural Networks (CNN). In this case, the CNN weights are updated following the FWI gradient, defining the process as a Physics-Informed Neural Network (PINN). FWI integrated with CNN has an important advantage. The CNN stabilizes the inversion, acting like a regularizer, avoiding local minima-related problems and sparing an initial velocity model in some cases. However, such a process, especially when not requiring an initial model, is computationally expensive due to the high number of iterations required until the convergence. In this work, we propose an approach which relies on combining supervised learning and physics-informed by using a previously trained CNN to start the DL-FWI inversion. Loading the pre-trained weights configures transfer learning. The pre-trained CNN is obtained using a supervised approach based on training with a reduced and simple data set to capture the main velocity trend at the initial FWI iterations. The proposed training process is different from the initial works on the area which obtained the velocity model from the shots in supervised learning tasks and that required a large amount of labelled data to ensure reasonable model predictions. We investigated in our approach two CNN architectures, obtaining more robust results and a reduced number of parameters when using a modified U-Net. The method was probed over three benchmark models, showing consistently that the pre-training phase reduces the process's uncertainties and accelerates the model convergence using minimal prior information. Besides, the final scores of the iterative process are better than the examples without transfer learning. Thus, transfer learning solved one main limitation of the previous PINN approaches: the unfeasible number of iterations when not using an initial model. Moreover, we tested the method using data with low-frequency band limitations, since the lack of low frequencies is a common issue within real seismic data. The inversion converges to reasonable results probing the method's robustness with restricted frequency content. [ABSTRACT FROM AUTHOR]

Accurate near-surface velocity models are necessary for land seismic imaging. First-arrival traveltime tomography (FTT) routinely used for estimating near-surface velocity models may fail in geological complex areas. Supervised deep learning (SDL) is capable of building accurate velocity models, based on tens of thousands of velocity model-shot gathers training pairs. It takes lots of time and memory space, which may be unaffordable for practical applications. We propose integrating the FTT and SDL to build near-surface velocity models. During the neural network training, the FTT-inverted models rather than the original seismic data are used as the network inputs and corresponding true models are the outputs. The FTT-inverted and true models are the same physical quantities and with the same dimensions. Their relationship is less non-linear than that between shot gathers and true models. Thus, the neural network of the proposed method can be trained well using only a small number of training samples, dramatically reducing the time and memory costs. Numerical tests demonstrate the feasibility and effectiveness of the proposed method. We applied the proposed method to a land data set obtained in mountainous areas in the west of China and obtained satisfactory near-surface velocity models and stacking images. [ABSTRACT FROM AUTHOR]

Extracting the dispersion energy image is one of the key steps of surface wave analysis, which is considered an effective and important way to obtain the vertical velocity structure. The quality of the dispersion energy image directly affects the accuracy of selected dispersion curves and the subsequent inversion. However, if we can theoretically obtain an accurate 2-D dispersion energy diagram, we will skip the step of extracting the dispersion curves. Here, we present the double-source frequency–wavenumber (DSFK) transform to extract theoretical 2-D dispersion energy images from multichannel data. Unlike 1-D dispersion energy images, which are based on the assumption of a horizontally layered medium structure, 2-D dispersion energy images also reflect information about the irregular interfaces of media. We adopt a geometry in which two reciprocal sources are on both ends of a receiver array. Then, the multichannel data obtained by this geometry are transformed into the frequency domain, and a 2-D dispersion energy image that corresponds to the 2-D structure is obtained by the DSFK transform. By producing a variety of synthetic data and processing active source data from the Salton Seismic Imaging Project, we confirm the accuracy and applicability of the DSFK transform. As demonstrated by the results, the DSFK transform can obtain the theoretical 2-D dispersion energy distribution and shows significant potential in the field of seismic tomography. [ABSTRACT FROM AUTHOR]

Seismic velocity analysis is the basis for seismic imaging and understanding complex subsurface geological structures. Although the performance of automatic velocity analysis methods based on Common Middle Point (CMP) data or Velocity Spectra (VS) is encouraging, particularly deep learning methods. However, most methods ignore the complementarity between CMP data and VS data, and only one of them is selected for velocity modelling. We propose a multimodal neural network (MMN) that combines the advantages of CMP data details representation and simplification of VS. MMN includes multilayer convolution structures and auto-encoder structures, which are used to extract time–space amplitude information from CMP gathers and energy groups features from VS data, respectively. This paper compared MMN with the CMP single-modal network (CSN) and the velocity spectra single-modal network (VSSN). Based on synthetic data, we investigated their differences in terms of continuity, accuracy, noise resistance and generalization. The MMN prediction results makes a trade-off between the overall continuity and local details. Visualization analysis of the intermediate feature maps explains the MMN velocity prediction mechanism, that is, the multi-angle representation and complementary fusion of velocity information. Finally, the performance of the proposed method is demonstrated using the braided river deposited field data example. [ABSTRACT FROM AUTHOR]

Due to complex geological conditions and instrument inaccuracy, raw seismic data are often characterized by low resolution. Deep learning is an emerging technique for seismic resolution improvement; however, its performance is often limited by the small amount of labeled data available. In this paper, we design a deep feature mining network (DFMN) to deal with this issue. DFMN has three components: shallow feature extraction block (SFB), deep feature mining block (DFB), and enhanced reconstruction block (ERB). First, the SFB component uses multi-scale kernels to learn rich information from low-resolution data. The convolutions incorporate the benefits of different kernel sizes, which are effective for shallow feature extraction. Second, the DFB component employs a dual-branch network architecture for deep feature mining. The dual-branch network learns more complementary features than a single-branch network, thus alleviating the requirement for large amounts of training data. Third, the ERB component combines the shuffled image and the interpolated image during reconstruction. Interpolated images, incorporating prior knowledge, can provide more contextual information in our model. The results show that DFMN is superior to a traditional upscaling algorithm and other deep learning methods in terms of (1) perceptual effects: more complete structural information, such as texture details; (2) quantitative evaluation indices: higher PSNR and SSIM; and (3) generalization ability: better performance on other data. [ABSTRACT FROM AUTHOR]

A major challenge for the oil and gas industry is expediting the exploration geophysics activity. It is necessary to produce fast and accurate image of the targeted hydrocarbon reservoir, yet it involves extensive dataset due to the target location lies in deeper subsurface. Therefore, target-oriented oil and gas-reservoir imaging is currently seen as more attractive than conventional full-volume migration, in terms of computation efficiency. Through the integration of seismic redatuming with Kirchhoff migration, we presented a target-oriented imaging workflow that can effectively reduce computational cost while simultaneously enhancing the accuracy of the migrated image. Kirchhoff migration is renowned for being one of the simplest and flexible migration techniques, making it straightforward to implement. However, this migration approach is typically avoided in the presence of complex geology, as it tends to introduce artifacts in the images, thereby disrupting the interpretation beneath the complex overburden. Therefore, this paper highlighted the advantages of utilizing seismic redatuming to retrieve data beneath the overburden, followed by the utilization of the redatumed data as input for Kirchhoff migration. With an improved signal to noise ratio, the primary reflection events were amplified, providing clearer representation of the subsurface structure within the hydrocarbon reservoir or targeted area. Additionally, we utilized the benefits of Kirchhoff migration and explored the feasibility of performing patch-based migration in several areas of interest beneath the salt layer. This approach aimed to reduce both the turnaround time and the associated seismic migration costs. The developed workflow then was subsequently put to the test and applied to both a sub-salt SEG Advanced Modeling (SEAM) model and field data situated in East Malaysia. Based on the outcomes of the target-oriented migration, the proposed methodology demonstrated its capability to generate better quality and instant migrated images within the designated region. Furthermore, the feasibility of target-oriented imaging with Kirchhoff migration was substantiated through the integration of seismic redatuming, resulting in significant benefits towards seismic processing, imaging, and interpretation. [ABSTRACT FROM AUTHOR]

The article focuses on exploring the architecture of magma plumbing systems and how magmas move through them in the context of volcanic activity. It discusses the use of advanced seismic techniques to study the shallowest parts of magmatic systems beneath the East Pacific Rise (EPR) and provides insights into the complex network of crustal magma bodies, shedding light on their characteristics and potential links to volcanic eruptions.

Seismic reflection interpretation at magma-poor rifted margins shows that crustal thinning within the hyper-extended domain occurs by in-sequence oceanward extensional faulting which terminates in a sub-horizontal reflector in the top-most mantle immediately beneath tilted crustal fault blocks. This sub-horizontal reflector is interpreted to be a detachment surface which develops sequentially with oceanward in-sequence crustal faulting. We investigate the geometry and evolution of active and inactive extensional faulting due to flexural isostatic rotation during magma-poor margin hyper-extension using a recursive adaptation of the rolling hinge model of Buck (1988) and compare modelling results with the seismic interpretation. In the case of progressive in-sequence faulting, we show that sub-horizontal reflectors imaged on seismic reflection data can be generated by the flexural isostatic rotation of faults with initially high-angle geometry. Flexural isostatic rotation produces shallowing of emergent fault angles, fault locking and the development of new high-angle short-cut fault segments within the hanging-wall. This results in the transfer and isostatic rotation of triangular pieces of hanging wall onto exhumed fault footwall, forming extensional allochthons which our modelling predicts are typically limited to a few km in lateral extent and thickness. While earthquake seismology favours a planar fault geometry with in the brittle seismogenic crust, seismic reflection imaging suggests a more listric geometry. Our modelling results show that a sequence of extensional listric or planar faults with identical parameters (i.e. location, heave, surface dip, Te) produce very similar sea-bed bathymetric relief. Listric and planar fault geometries do however produce distinct Moho and allochthon shapes. We propose that the initial fault geometry, prior to flexural isostatic rotation, is planar in the seismogenic crust becoming listric at depth as the brittle plastic transition is approached. Extensional faulting and thinning of hyper-extended continental crust may eventually lead to mantle exhumation. Where extensional faulting is in-sequence, this results in a smooth bathymetric transition from thinned continental crust to exhumed mantle. In contrast out-of-sequence faulting results in a transition to exhumed mantle with bathymetric relief. [ABSTRACT FROM AUTHOR]

The viscoelasticity of an underground medium will cause absorption and attenuation of seismic waves, resulting in energy attenuation and phase distortion. This absorption and attenuation is often quantified by the quality Factor Q. The strong attenuation effect resulting from geology is a challenging problem for high-resolution imaging. To compensate for the attenuation effect, it is necessary to estimate the attenuation parameters accurately. However, it is difficult to directly derive a heterogeneous attenuation Q model. This research letter proposes a method to derive a Q model from reflection seismic data using a backpropagation neural network (BPNN), one of the most widely used neural network models. We treated the Q detection problem as a pattern recognition task and train a network to assign the correct Q classes to a set of input patterns. The proposed method uses synthetic data for network training and validation. Finally, we used a set of model data and a set of field data to demonstrate the effectiveness of this method, and the high-resolution imaging results in the time domain with appropriate compensation are obtained. [ABSTRACT FROM AUTHOR]

The Java ‐ Lesser Sunda margin, which features multi‐scale subducting oceanic basement relief, is classified as neutral (Lombok and Sumbawa) to erosional (Central Java to Bali) in comparison to its accretionary counterpart offshore Sumatra. However, a comprehensive analysis of how plate boundary and upper plate structure across the neutral to erosional transition are modulated by the subduction of oceanic basement relief is lacking to date. To shed light on the tectonic parameters that push the margin into the neutral or erosional domain, we combine multi‐channel reflection seismic images derived through a grid‐based P‐wave velocity inversion, and multibeam bathymetric maps. The data document how different scales of subducting topography modify seafloor morphology, upper plate structure, and décollement position. Large‐scale subducting features cause a landward shift of the deformation front, shortening of the accretionary wedge, and seafloor steepening at the relief's trailing edge. Small‐scale subducting ridges primarily impact the frontal prism resulting in over‐steepening at the trench and localized slope failure. Ahead of subducting relief, deformation of the accretionary wedge encompasses enhanced compression and a reduction in seafloor slope but appears independent of the size of the relief. Ridge and seamount subduction induce frontal erosion and basal erosion offshore Lombok and Bali, respectively. Our P‐wave velocity models indicate that the rigidity of the upper plate's base along the eastern Sunda margin is significantly lower than the worldwide trend. We conclude that this favors the genesis of tsunami earthquakes that have occurred on the Java margin. Plain Language Summary: The convergence of the tectonic plates drives a wide variety of geological processes along the plate margins, including the formation of the forearc accretionary wedge, volcanic activities, and megathrust earthquakes. Over the past 40 years, marine research shows that different sizes of oceanic reliefs (seamounts and ridges) are widely distributed over the seafloor, approaching the trench, and eventually subducted underneath the overriding plate. An accurate observation of the subducted reliefs and their tectonic impact on the overriding plate depends on different observation approaches, data processing methods, and the evolutionary history of the forearc. In the Java margin, the oceanic seafloor features massive seamounts with different scales and shapes, and the bathymetry of the overriding plate is highly disturbed. This provides us with the best opportunity of studying the rugged seafloor's seismogenic and geological impacts. By using state‐of‐the‐art seismic imaging techniques, we image the subsurface structures, obtain the forearc velocity, identify the seamounts, and discuss the seamounts' effect on structural deformation and megathrust earthquake occurrence. Distinctively, the marine forearc gets shortened and thickened significantly by seamount subduction. Structural images sharply illustrate different deformation patterns and stress regimes at the seamount's different flanks and reveal the possible process of subduction erosion. Key Points: Upper plate deformation scales with variable subducting relief, as observed along the Java Trench in seismic sections and bathymetrySubduction of seafloor topography induces progression from an accretion‐dominated domain toward a phase of subduction erosionThe overall low rigidity of the upper plate's base may contribute to the Java margin earthquake's tsunami‐genesis [ABSTRACT FROM AUTHOR]

The mantle transition zone (MTZ) plays an important role in modulating material transport between the upper mantle and the lower mantle. Constraining this transport is essential for understanding geochemical reservoirs, hydration cycles, and the evolution of the Earth. Slabs and hotspots are assumed to be the dominant locations of transport. However, the degree of material transport in other areas is debated. We apply P‐to‐S receiver functions to an amphibious data set from Cascadia to image the MTZ discontinuities beneath mid‐ocean ridges, a hotspot, and a subduction zone. We find a MTZ thinned by 10 ± 6 km beneath the ridges and by 8 ± 4 km beneath the base of the slab, closely resembling the 660 discontinuity topography. Depressions on the 410 discontinuity are smaller, 5 ± 2 km on average, focused in the north and the south and accompanied by supra‐410 discontinuity melt phases. The depressions occur away from locations of uplifted 660 discontinuity, but near slow seismic velocity anomalies imaged in the upper mantle. This suggests lower mantle upwellings occur beneath ridges and beneath the base of slabs but stall in the transition zone, with upper mantle convection determining upward material transport from the transition zone. Therefore, upper mantle dynamics play a larger role in determining transfer than typically assumed. Plain Language Summary: The mantle transition zone (MTZ), the region between the upper mantle and lower mantle, plays an important role in material transport during mantle convection. Subduction zones and hotspots are regarded as the dominant locations where material moves across the MTZ, but it is debatable in other areas. We apply P‐to‐S receiver functions which reflect the inner seismic discontinuities of the Earth to a seismic data set from Cascadia. We find the MTZ is thinned beneath the ridges and beneath of the base of slabs, closely resembling the 660 discontinuity topography, likely caused by upwellings from the lower mantle. Depressions on the 410 discontinuity, inferred to be upwellings to the upper mantle, are smaller and occur away from locations of upwelling indicated by the uplifted 660. However, they are near slow seismic velocity anomalies imaged in the upper mantle. Lower mantle upwellings occur beneath ridges and beneath the base of slabs but instead of ascending vertically as typical diagrams demonstrate they may stall in the transition zone and be guided by the upper mantle convection. The results suggest upwellings may take place in regions other than hotspots, and upper mantle dynamics play a larger role in determining transfer than typically assumed. Key Points: P‐to‐S receiver functions image mantle transition zone thinning and an uplifted 660 beneath the Cascadia Ridges and behind the slabsDepressions of the 410 are offset from the uplifted 660s but in agreement with low seismic velocity anomalies of the upper mantleThis suggests upper mantle dynamics dictate vertical flow from the shallow transition zone including entrained hydration [ABSTRACT FROM AUTHOR]

Gas hydrates that occur on many continental margins have received global attention. In reflection seismic imaging, the bottom-simulating reflector (BSR) is a common indicator of gas hydrates. However, it is difficult to identify gas hydrates and quantify their amounts through the BSR alone. For gas-hydrate characterization, it is therefore useful to measure seismic stratigraphic and attenuation attributes. Short-scale patterns of layering that contain information about the amount and mechanism of gas hydrates can be identified through stratigraphic and attenuation attributes. We measure the complete time-variant spectra by using sparse strongest peaks, and the spectral differences at different times through attenuation parameters Q –1 and γ. The traditional Q –1 is associated with the attenuation of the frequency-dependent part of wavefield, and the γ characterizes the frequency-independent attenuation. The measurement approach is straightforward and requires no sophisticated inverse algorithm and is applied to surface seismic data acquired over the Hikurangi and Gondwana margins, eastern New Zealand. High-quality spectral and attenuation images are obtained. Spectral attributes correlate with BSRs and large positive Q –1 and negative γ -values are below and above the BSRs, which are interpreted as being related to free-gas and gas-hydrate accumulations. These results will aid the quantification of gas hydrates and the assessment of their roles as an energy resource, as a potential geological hazard, and in climate change and ocean warming. [ABSTRACT FROM AUTHOR]

We present an integrated method for mapping the basement structures of sedimentary basins by combining surface gravity data, seismic imaging, and borehole logging information. The core of the method is a nonlinear inversion algorithm for constructing the shape and depth of the basement from surface gravity data. By using the primal-logarithmic barrier method, we impose depth constraints from the borehole information. The basement depth was imaged by seismic interpretation and incorporated into the inversion as a reference model. As a result, the gravity inversion constructs basement structures that are closest to the seismic input while simultaneously satisfying the surface gravity data and borehole information. We used this new methodology to unveil the basement morphology of the Recôncavo Basin, Brazil. Recôncavo is a syn-rift onshore mature basin that exhibits a strong correlation between oil field distribution and tectonic framework. The seismic imaging in the area is ambiguous, and our approach improved the basement definition and highlighted exploration targets in the studied area. [ABSTRACT FROM AUTHOR]

Fault detection in seismic images is important in interpreting seismic changes because the faults confirm unexpected subsurface topographical changes. In the traditional approach, faults are manually detected from post-stack seismic data and identified as reflection discontinuities, which is a very tedious process. To increase efficiency and reduce time consumption, a number of automatic fault detection techniques have been developed, of which techniques based on deep learning have proven to be efficient. This work proposed a deep learning based method to generate a seismic attribute called fault probability to highlight fault zones in the 3D seismic images. Since the technique detects the faults directly from seismic volumes, pre-computed attributes such as those used in coherence or edge detection methods are not strictly necessary. The proposed technique is instigated in two stages--training and prediction. In the training stage, a CNN model is trained with real data taken from 7 annotated seismic volumes, in which every point is labeled as fault or no-fault. Then in the prediction stage, the trained network is used to calculate the fault probability at every location in the new seismic image volumes. Both synthetic and real data sets are used to validate the proposed method. The obtained results proved that the proposed deep learning-based fault detection method outperforms some existing methods and also achieves effective performance compared to humans on an expert labelled seismic image dataset, much exactly predicting subtle faults which cannot annotate by an expert interpreter. [ABSTRACT FROM AUTHOR]

In this study, we explore the latest generation of seismic imaging algorithms (migration) that have been successful in the oil and gas exploration industry and apply them in the more challenging hard rock environment. Seismic migration is a crucial processing step required to build an accurate image of the Earth's subsurface. The seismic method applied in a hard rock environment has a specific set of challenges: complex geological settings often comprised of steeply dipping interfaces and heterogeneities (faults, fracture zones, thrusts, etc.); spatially variable zones of alteration, low intrinsic signal‐to‐noise ratio; complex near‐surface conditions (the weathered overburden has a very high contrast in seismic properties with base formations). Here, we present how a dense source–receiver 3D grid in combination with the latest pre‐stack depth imaging techniques can image the geology with a remarkable level of detail and to depths as shallow as the top of fresh rock. We also propose a comprehensive velocity model building strategy applied specifically to the Carosue Dam deposit, Western Australia, which is mainly characterized by volcaniclastic and volcanic rocks within the Carosue Basin disrupted by a complex system of gold‐bearing faults. To arrive at an optimized velocity model for pre‐stack depth imaging, we combine drillhole data with tomographic refinement. Excellent correlation between resultant seismic images and refraction tomography, vertical seismic profiling and drillholes can be attributed to the integrated velocity model building workflow used for depth migration. [ABSTRACT FROM AUTHOR]