Controlled Source Seismology Research Articles

Cretaceous magmatic underplating and delamination beneath continental SE Brazil and their tectonic implications: Evidence from the PABBRISE wide-angle reflection and refraction seismic profile

Controlled-source seismology provides robust methods to obtain information on the velocity structure of the crust and uppermost mantle. We present the first systematic wide-angle reflection and refraction (WARR) seismic experiment for continental southeastern Brazil: the NW-SE-oriented and ca. 700 km-long PABBRISE (PAraná Basin, Brasília and RIbeira orogens Seismic Experiment) profile. Using ray tracing, we obtained a two-dimensional P-wave seismic velocity model of the crust and uppermost mantle. Our model shows that Moho depths vary between ca. 31 and 45 km, with crustal thinning towards the continental margin. The modeled velocity structure sheds new light on the present-day lithospheric morphology of the region, with the most striking features noted within the lower crust and uppermost mantle. We show that beneath the Paraná Basin, parts of the lower crust and uppermost mantle were pervasively intruded due to magmatic underplating. We modeled a ca. 200 km-long and 11 km-thick (maximum thickness), high-velocity (7.20–7.80 km/s) lower crustal (HVLC) body in this region. Outside the Paraná Basin, near the continental margin, comparatively thinner crust (∼31 km) with low mean P-wave velocity (6.20 km/s) and a low velocity (<7.90 km/s) uppermost mantle probably are the result of delamination. Magmatic underplating and delamination can be linked with the breakup of Pangea and the opening of the South Atlantic Ocean during the Cretaceous.

Test of P-wave receiver functions for a seismic velocity and gravity model across the Baikal Rift Zone

SUMMARY The seismic receiver function (RF) technique is widely used as an economic method to image earth's deep interior in a large number of seismic experiments. P-wave receiver functions (RFs) constrain crustal thickness and average Vp/Vs in the crust by analysis of the Ps phase and multiples (reflected/converted waves) from the Moho. Regional studies often show significant differences between the Moho depth constrained by RF and by reflection/refraction methods. We compare the results from RF and controlled source seismology for the Baikal Rift Zone by calculating 1480 synthetic RFs for a seismic refraction/reflection velocity model and processing them with two common RF techniques [H–κ and Common Conversion Point (CCP) stacking]. We compare the resulting synthetic RF structure with the velocity model, a density model (derived from gravity and the velocity model), and with observed RFs. Our results demonstrate that the use of different frequency filters, the presence of complex phases from sediments and gradual changes in the properties of crustal layers can lead to erroneous interpretation of RFs and incorrect geological interpretations. We suggest that the interpretation of RFs should be combined with other geophysical methods, in particular in complex tectonic regions and that the long-wavelength Bouguer gravity anomaly signal may provide effective calibration for the determination of the correct Moho depth from RF results. We propose and validate a new automated, efficient method for this calibration.

An extension of least-squares redatuming: Simultaneous reconstruction of overburden reflectivities and virtual data

In a controlled-source seismology scenario, scattered wavefields from deep targets, e.g., the petroleum reservoirs, are often distorted by strong heterogeneities of the overburden media, complex topography and irregular acquisition. It is key for target-oriented seismic imaging to accurately reconstruct the scattering wavefields from the deep structures. Recent studies have promoted seismic redatuming for removing the imprint of the overburden through reconstructing virtual data just above the target area. Among the model-based redatuming approaches, least-squares inversion has advantages in mitigating the artifacts due to limited recording aperture and irregular spatial sampling, and allowing for the introduction of appropriate constraints to the reconstructed data. However, conventional least-squares redatuming (LSR) suffers from the spurious events and noises associated with the wavefield interaction of the overburden reflectivities close to the datum. To overcome this problem, we establish an efficient double-parameter LSR method that simultaneously updates the overburden reflectivities and the virtual data through a constrained optimization. Synthetic and field data examples demonstrate that this extended LSR method can effectively retrieve the scattering wavefields from the deep part and improve the target-oriented prestack depth migration below the datum.

Automatic Image-Based Event Detection for Large-N Seismic Arrays Using a Convolutional Neural Network

Passive seismic experiments have been proposed as a cost-effective and non-invasive alternative to controlled-source seismology, allowing body–wave reflections based on seismic interferometry principles to be retrieved. However, from the huge volume of the recorded ambient noise, only selected time periods (noise panels) are contributing constructively to the retrieval of reflections. We address the issue of automatic scanning of ambient noise data recorded by a large-N array in search of body–wave energy (body–wave events) utilizing a convolutional neural network (CNN). It consists of computing first both amplitude and frequency attribute values at each receiver station for all divided portions of the recorded signal (noise panels). The created 2-D attribute maps are then converted to images and used to extract spatial and temporal patterns associated with the body–wave energy present in the data to build binary CNN-based classifiers. The ensemble of two multi-headed CNN models trained separately on the frequency and amplitude attribute maps demonstrates better generalization ability than each of its participating networks. We also compare the prediction performance of our deep learning (DL) framework with a conventional machine learning (ML) algorithm called XGBoost. The DL-based solution applied to 240 h of ambient seismic noise data recorded by the Kylylahti array in Finland demonstrates high detection accuracy and the superiority over the ML-based one. The ensemble of CNN-based models managed to find almost three times more verified body–wave events in the full unlabelled dataset than it was provided at the training stage. Moreover, the high-level abstraction features extracted at the deeper convolution layers can be used to perform unsupervised clustering of the classified panels with respect to their visual characteristics.

Random noise attenuation of sparker seismic oceanography data with machine learning

Abstract. Seismic oceanography (SO) acquires water column reflections using controlled source seismology and provides high lateral resolution that enables the tracking of the thermohaline structure of the oceans. Most SO studies obtain data using air guns, which can produce acoustic energy below 100 Hz bandwidth, with vertical resolution of approximately 10 m or more. For higher-frequency bands, with vertical resolution ranging from several centimeters to several meters, a smaller, low-cost seismic exploration system may be used, such as a sparker source with central frequencies of 250 Hz or higher. However, the sparker source has a relatively low energy compared to air guns and consequently produces data with a lower signal-to-noise (S∕N) ratio. To attenuate the random noise and extract reliable signal from the low S∕N ratio of sparker SO data without distorting the true shape and amplitude of water column reflections, we applied machine learning. Specifically, we used a denoising convolutional neural network (DnCNN) that efficiently suppresses random noise in a natural image. One of the most important factors of machine learning is the generation of an appropriate training dataset. We generated two different training datasets using synthetic and field data. Models trained with the different training datasets were applied to the test data, and the denoised results were quantitatively compared. To demonstrate the technique, the trained models were applied to an SO sparker seismic dataset acquired in the Ulleung Basin, East Sea (Sea of Japan), and the denoised seismic sections were evaluated. The results show that machine learning can successfully attenuate the random noise in sparker water column seismic reflection data.

Seismology in India

The controlled source seismology (CSS), where artificial sources are used and near-vertical reflections, refractions and wide-angle reflections are recorded, has been a powerful tool for delineating shallow/deep crustal structures and subsurface features. Significant results, brought out from the vintage and new CSS data in the 21st century, include geological/tectonic aspects of the Dharwar craton, Delhi-Aravalli fold belt, Central Indian Tectonic Zone, Southern Granulite Terrain, Proterozoic Cuddapah basin, Kangra fold-thrust belt, Hazara syntaxis in NW Himalaya, Chambal-valley Vindhya basin, Mahanadi delta and Bengal basins. These studies also provide useful inputs in understanding the genesis of 1997 Jabalpur and 2001 Bhuj earthquakes. Imaging sub-volcanic Mesozoic sediments, in which more than 50% of global oil is found, has been a challenge by routine geophysical methods. This problem has been alleviated by CSS experiments, and Mesozoic sediments below the Deccan volcanic rocks have been delineated in the Saurashtra peninsula, Kutch peninsula, Tapti-graben and Kerala-Konkan offshore regions, and Gondwana sediments below the Rajmahal Traps have been mapped in the Mahanadi and Bengal basin. Several approaches have been proposed for the detection, characterization and assessment of gashydrates from shallow seismic data. These have been applied to field data that has resulted into identifying gas-hydrates, a major future energy resource of India, in the Krishna-Godavari, Mahanadi and Andaman offshore regions. Gas-hydrates have been subsequently recovered by drilling & coring under India National Gas Hydrates Program. The prognosticated amount of methane stored as gas-hydrates within Indian Exclusive Economic Zone is so huge that only 10% production may suffice our vast energy requirement for the next 100 years.

Controlled Source Seismology in India in the 21st Century

Controlled Source Seismology in India in the 21st Century

An approach of refraction seismology for processing and modeling of local earthquake seismogram sections of virtual sources at multiple depths in seismogenic regions—application to Koyna-Warna region, India, for upper crustal P and S velocity structure

We present an approach based on controlled source seismology (CSS) methods, especially developed for processing and modeling of the local earthquake seismograms. Record sections of the local earthquake seismograms generated for multiple source depths illuminate the upper crustal velocity structure in the region. Extensive travel times and synthetic seismograms modeling of the observed record sections reveal the P and S velocity structure in the region. The strength of this approach essentially lies with the possibility of validating the upper crustal velocity models inferred in various subregions of the seismogenic region. A redundant and significantly large number of virtual source local earthquake seismogram sections, gathered for multiple source depths and varying source mechanisms in each of the subregions, validate the same set of P and S velocity models in that region. Further, those models are found to generate the synthetic seismograms consistent with the observed sections. The proposed approach effectively utilizes a reliable dataset from a great volume of well-located local earthquake recordings of a state-of-the-art digital seismograph network. Such a dataset of local earthquake seismograms in the Koyna-Warna active earthquake zone is used here to demonstrate this approach and obtained subregion-specific models of upper crustal P and S velocity structure in the epicentral region. The results indicate that the technique presented here is efficient for processing and modeling the local earthquake seismograms and deriving upper crustal velocity models in the seismogenic regions.

High-resolution surface wave tomography of the European crust and uppermost mantle from ambient seismic noise

Taking advantage of the large number of seismic stations installed in Europe, in particular in the greater Alpine region with the AlpArray experiment, we derive a new high-resolution 3-D shear wave velocity model of the European crust and uppermost mantle from ambient-noise tomography. The correlation of up to 4 yr of continuous vertical-component seismic recordings from 1293 broad-band stations (10 • W-35 • E, 30 • N-75 • N) provides Rayleigh wave group velocity dispersion data in the period band 5-150 s at more than 0.8 million virtual source-receiver pairs. 2-D Rayleigh wave group velocity maps are estimated using adaptive parametrization to accommodate the strong heterogeneity of path coverage. A probabilistic 3-D shear wave velocity model, including probability densities for the depth of layer boundaries and S-wave velocity values, is obtained by nonlinear Bayesian inversion. A weighted average of the probabilistic model is then used as starting model for the linear inversion step, providing the final V s model. The resulting S-wave velocity model and Moho depth are validated by comparison with previous geophysical studies. Although surface wave tomography is weakly sensitive to layer boundaries, vertical cross-sections through our V s model and the associated probability of the presence of interfaces display striking similarities with reference controlled-source seismology (CSS) and receiver function sections across the Alpine belt. Our model even provides new structural information such as an ∼8 km Moho jump along the CSS ECORS-CROP profile that was not imaged by the reflection data due to poor penetration across a heterogeneous upper crust. Our probabilistic and final shear wave velocity models have the potential to become new reference models of the European crust, both for crustal structure probing and geophysical studies including waveform modelling or full-waveform inversion.

Crustal structure and evolution of the Arctic Caledonides: Results from controlled-source seismology

The continuation of the Caledonides into the Barents Sea has long been a subject of discussion, and two major orientations of the Caledonian deformation fronts have been suggested: NNW-SSE striking and NE-SW striking. A regional NW-SE oriented ocean bottom seismic profile across the western Barents Sea was acquired in 2014. In this paper we map the crust and upper mantle structure along this profile in order to discriminate between different interpretations of Caledonian structural trends and orientation of rift basins in the western Barents Sea. Modeling of P-wave travel times has been done using a ray-tracing method, and combined with gravity modeling. The results show high P-wave velocities (4km/s) close to the seafloor, as well as localized sub-horizontal high velocity zones (6.0km/s and 6.9km/s) at shallow depths which are interpreted as magmatic sills. Refractions from the top of the crystalline basement together with reflections from the Moho give basement velocities from 6.0km/s at the top to 6.7km/s at the base of the crust. P-wave travel time modeling of the OBS profile indicate an eastwards increase in velocities from 6.4km/s to 6.7km/s at the base of the crystalline crust, and the western part of the profile is characterized by a higher seismic reflectivity than the eastern part. This change in seismic character is consistent with observations from vintage reflection seismic data and is interpreted as a Caledonian suture extending through the Barents Sea, separating Barentsia and Baltica. Local deepening of Moho (from 27km to 33km depth) creates “root structures” that can be linked to the Caledonian compressional deformation or a suture zone imprinted in the lower crust. Our model supports a separate NE-SW Caledonian trend extending into the central Barents Sea, branching off from the northerly trending Svalbard Caledonides, implying the existence of Barentsia as an independent microcontinent between Laurentia and Baltica.

Crustal Structure of the Tengchong Intra-plate Volcanic Area

We here provide an overview of our current understanding of the crustal structure of Tengchong in southwest China, a key intra-plate volcanic area along the Himalayan geothermal belt. Given that there is hitherto a lack of information about the near-surface structure of intra-plate volcanic areas, we present the first seismic reflection and velocity constraints on the shallow crust between intra-plate volcanoes. Our near-surface seismic images reveal the existence of dome-shaped seismic reflectors (DSRs) in the shallow crust between intra-plate volcanic clusters in Tengchong. The two DSRs are both ~2 km wide, and the shallowest parts of the DSRs are found at the depth of 200–300 m. The velocity model shows that the shallow low-velocity layer (<4 km/s) is anomalously thick (~1 km) in the region where the DSRs are observed. The presence of DSRs indicates significant levels of intra-plate magmatism beneath the along-axis gap separating two volcano clusters. Along-axis gaps between volcano clusters are therefore not necessarily an indicator of lower levels of magmatism. The seismic images obtained in this technically challenging area for controlled-source seismology allow us to conclude that shallow crustal structures are crucial for understanding the along-axis variations of magmatism and hydrothermal activities in intra-plate volcanic areas.

The lithosphere and asthenosphere system in Italy as inferred from the Vp and Vs 3D velocity model and Moho map

We present an updated high-resolution tomographic P- and S-wave velocity model of the lithosphere and asthenosphere system in Italy, obtained by adding the observations from ∼25,000 earthquakes (recorded between 2003 and April 2012 at three-component stations) to the previously inverted dataset (165,000 P-wave phases from ∼8000 events recorded between 1988 and 2002 at single-component). The final dataset includes ∼586,000 P- and ∼290,000 S-wave arrival times. The main strength of this research is the use of a method able to model P- and S-seismic phases refracted at the Moho discontinuity. We use a new and original map of the 3D Moho geometry obtained by integrating selected high quality controlled source seismology (CSS) and teleseismic receiver function data. Resolution strongly benefits also from the fast increase in number and quality of INGV National Seismic Network since year 2003 and from its integration with several permanent regional seismic networks. This study confirms the main structural features in the best-resolved parts of the inverted volume and much better images details in the less resolved areas of previous Vp models, due to both the larger number of inverted phases and the more even distribution of seismic stations. A new 3D Vp/Vs model is presented and discussed. Shallow structures and the relationships between the Adriatic, Tyrrhenian, and Adria plates are better imaged.

Models of the Earth’s crust from controlled-source seismology — Where we stand and where we go?

Controlled-source seismology (CSS) is the primary source of information regarding the fine structure of the lithosphere. The aim of this paper is to provide an overview of the methods that are commonly used to derive Earth models from CSS data with the focus on the wide-angle reflection/refraction method. Some outlook on the future of the CSS is presented with the special emphasis on the full-wavefield based methods like full-waveform inversion, which brings high level of objectivity into modeling, as well as significantly increases spatial resolution. It is stressed that the researchers should be aware of the limitations of how the elastic parameters transcribe into the actual rock properties which should stimulate them to go beyond the simple P-wave modeling and to build multiparameter Earth models based either on the seismic data or constrained by additional geophysical fields in order to derive sound geological interpretation of their models.

100 years of seismic research on the Moho

Abstract The detection of a seismic boundary, the “Moho”, between the outermost shell of the Earth, the Earth's crust, and the Earth's mantle by A. Mohorovicic was the consequence of increased insight into the propagation of seismic waves caused by earthquakes. This short history of seismic research on the Moho is primarily based on the comprehensive overview of the worldwide history of seismological studies of the Earth's crust using controlled sources from 1850 to 2005, by Prodehl and Mooney (2012) . Though the art of applying explosions, so-called “artificial events”, as energy sources for studies of the uppermost crustal layers began in the early 1900s, its effective use for studying the entire crust only began at the end of World War II. From 1945 onwards, controlled-source seismology has been the major approach to study details of the crust and underlying crust–mantle boundary, the Moho. The subsequent description of history of controlled-source crustal seismology and its seminal results is subdivided into separate chapters for each decade, highlighting the major advances achieved during that decade in terms of data acquisition, processing technology, and interpretation methods. Since the late 1980s, passive seismology using distant earthquakes has played an increasingly important role in studies of crustal structure. The receiver function technique exploiting conversions between P and SV waves at discontinuities in seismic wavespeed below a seismic station has been extensively applied to the increasing numbers of permanent and portable broad-band seismic stations across the globe. Receiver function studies supplement controlled source work with improved geographic coverage and now make a significant contribution to knowledge of the nature of the crust and the depth to Moho.

Combining controlled-source seismology and receiver function information to derive 3-D Moho topography for Italy

Combining controlled-source seismology and receiver function information to derive 3-D Moho topography for Italy

A parametric analysis of two-dimensional elastic full waveform inversion of teleseismic data for lithospheric imaging

The development of dense networks of broad-band seismographs makes teleseismic data amenable to full-waveform inversion (FWI) methods for high-resolution lithospheric imaging. Compared to scattered-field migration, FWI seeks to involve the full seismic wavefield in the inversion. We present a parametric analysis of 2-D frequency-domain FWI in the framework of lithospheric imaging from teleseismic data to identify the main factors that impact on the quality of the reconstructed compressional (P)-wave and shear (S)-wave speed models. Compared to controlled-source seismology, the main adaptation of FWI to teleseismic configuration consists of the implementation with a scattered-filed formulation of plane-wave sources that impinge on the base of the lithospheric target located below the receiver network at an arbitrary incidence angle. Seismic modelling is performed with a hp-adaptive discontinuous Galerkin method on unstructured triangular mesh. A quasi-Newton inversion algorithm provides an approximate accounting for the Hessian operator, which contributes to reduce the footprint of the coarse acquisition geometry in the imaging. A versatile algorithm to compute the gradient of the misfit function with the adjoint-state method allows for abstraction between the forward-problem operators and the meshes that are during seismic modelling and inversion, respectively. An approximate correction for obliquity is derived for future application to real teleseismic data under the two-dimension approximation. Comparisons between the characteristic scales involved in exploration geophysics and in teleseismic seismology suggest that the resolution gain provided by full waveform technologies should be of the same order of magnitude for both applications. We first show the importance of the surface-reflected wavefield to dramatically improve the resolving power of FWI by combining tomography-like and migration-like imaging through the incorporation of the forward-scattered and the backscattered wavefields in the inversion. The resolution of FWI is assessed through checkerboard tests and confirms a resolution of the order of the wavelength for both the P and S speeds, when the full wavefield is incorporated in the inversion. Secondly, we show that computationally efficient strategies, which consist of decimating the number of frequency components involved in the inversion, do not apply to teleseismic acquisitions, because the scattering-angle bandwidth sampled by plane-wave sources can be narrow and coarsely sampled, compared to that provided by dense profiles of point sources in exploration seismology. The waveform inversion is less sensitive to the band of incidence angles spanned by the plane-wave sources and to the sampling of this band. However, the deficit of vertically propagating plane waves hampers the vertical resolution of planar layers. Aliasing artefacts created by coarse arrays of receivers are illustrated. We show how taking into account the Hessian in the inversion and the suitable management of frequencies in the inversion help to mitigate these artefacts. Acceptable reconstructions are shown for both the P- and S-wave speeds for a receiver spacing of up to 20 km in the 0.1–0.4 Hz frequency range. Building a reliable initial model for FWI is a highly non-linear problem in exploration seismology. We show how the low-frequency content of the teleseismic sources allow us to build accurate P- and S-wave speed models starting from simple vertical-gradient velocity models. All of these results are derived using a 2-D realistic synthetic experiment that was performed with noise-free data. Most of conclusions of this parametric study should apply in three dimensions. The impact of noise and the footprint of other experimental parameters such as the estimation of the temporal plane-wave signature and the estimation of the incidence angle of the impinging plane waves, need however to be assessed in the future.

High-precision earthquake locations in Switzerland using regional secondary arrivals in a 3-D velocity model

We present a new approach to relocate earthquakes in the greater western Alpine region using main crustal phases (Pg, Pn, PmP) that takes advantage of recent developments in P-wave velocity models and modelling of the Moho topography in the region, as well as the ability to track reflected and refracted phases in three-dimensional (3-D) heterogeneous media. Our approach includes a new 3-D P-wave velocity model for Switzerland and surrounding regions that combines a first-order Moho discontinuity based on local earthquake tomography (LET) and controlled-source seismology (CSS) information and 3-D seismic velocity information based on LET. Traveltimes for the main crustal phases (Pg, Pn, PmP) are computed using a fast marching method. We use a non-linear, probabilistic approach to relocate earthquakes that has been extended to include the use of secondary phases. We validate our approach using synthetic data, which was computed for a real earthquake and different combinations of available phases (Pg, Pn, PmP). We also applied our approach to relocate four selected earthquakes, two shallow and two deep crustal events in the northern Alpine foreland, for which independent information (ground truth information) on their focal depths exist. Our results demonstrate that the precision and accuracy of focal depth estimates can be greatly improved if secondary phases are used. This gain is a combined effect of an improved range of take-off angles and the use of differential traveltimes between first and secondary arriving phases. Our results also show that reliable information on the Moho depth is crucial to obtain accurate focal depths, if Pn or PmP phases are used in the relocation process. Finally, our approach demonstrates that proper identification of the main crustal phases in combination with an appropriate model parametrization in the forward solver will significantly improve earthquake locations.

Combining controlled-source seismology and local earthquake tomography to derive a 3-D crustal model of the western Alpine region

SUMMARY We present a newly developed approach of combining controlled-source seismology (CSS) and local earthquake tomography (LET) data to obtain a new 3-D crustal model of the western Alpine region. Our approach combines either data by taking into account the strengths of the individual seismic methods. Our western Alpine 3-D model is primarily based on a welldefined Moho, constrained by CSS and LET data, and includes smooth lateral variations in seismic velocities mainly constrained by LET data, but locally also by CSS data. The consistent combination of results from the two different seismic methods is feasible due to LET Moho elements, as defined by characteristic P-wave velocities and their uncertainty estimates. These uncertainty estimates are based on values of the diagonal element of the resolution matrix, absolute P-wave velocities that are typical for crust and mantle and a specific velocity gradient across the Moho discontinuity. Finally, our definition of LET Moho elements and their uncertainties is validated by comparisons of highest quality Moho results from both methods coinciding in 353 localities. Our model clearly shows three Moho surfaces, being Europe, Adria and Liguria as well as major tectonic structures like suture zones and the high-velocity Ivrea body. In general, it is in a good agreement with previous studies. The biggest differences occur along plate boundaries, where the strong lateral velocity variations arebestresolvedbyLET.DuetothelargernumberofavailableMohoreflectorelementsamore accurate definition of plate boundaries at Moho level is possible and, therefore, new insights in deep lithosphere structures of the Alpine collision zone can be expected. Furthermore, our new 3-D crustal model directly includes a 3-D migrated image of the Ivrea body.

Bayesian Strategies to Assess Uncertainty in Velocity Models

Quantifying uncertainty in models derived from observed seismic data is a major issue. In this research we examine the geological structure of the sub-surface using controlled source seismology which gives the data in time and the distance between the acoustic source and the receiver. Inversion tools exist to map these data into a depth model, but a full exploration of the uncertainty of the model is rarely done because robust strategies do not exist for large non-linear complex systems. There are two principal sources of uncertainty: the first comes from the input data which is noisy and band-limited; the second is from the model parameterisation and forward algorithm which approximate the physics to make the problem tractable. To address these issues we propose a Bayesian approach using the Metropolis-Hastings algorithm.

Correction to “Three-dimensional Moho topography in Italy: New constraints from receiver functions and controlled source seismology”

Correction to “Three-dimensional Moho topography in Italy: New constraints from receiver functions and controlled source seismology”