Ambient Seismic Field Research Articles

Seismic attenuation tomography of Eastern Europe from ambient seismic noise analysis

SUMMARY The Eastern-Europe region (EER), is a complex geotectonic area that captures part of the Alpine-Himalayan Orogen, the subduction of multiple NeoTethys Branches and part of the East European Craton. It is one of the most exciting geological areas in Europe due to a diversity of tectonic processes acting within it: extensional basin evolution, oceanic subduction, post-collisional volcanism, as well as active crustal deformation associated with the push of the Adria plate or the pull of the actively detaching Vrancea slab. This makes EER an excellent natural laboratory to study the behaviour of the lithosphere–asthenosphere system in a heterogeneous tectonic setting. To investigate the lateral heterogeneity and physical properties of the crust in the EER, we use ambient seismic noise data recorded by the vertical components of broad-band stations that have been operational between 1999 and 2020 in Eastern Europe and surrounding regions. We used this significant amount of data and the latest processing techniques of the ambient seismic noise field based on the continuous wavelet transform to compute cross-correlations between various station pairs, turning every available seismic station into a virtual source. The coda of the interstation cross-correlograms were used to determine coda quality factors (Qc) of Rayleigh waves in four different period ranges (3.0–5.0, 5.0–10.0, 10.0–20.0 and 20.0–30.0 s) and to invert them in the 2-D space, constructing the highest resolution attenuation tomography of the region. Our results reveal high attenuation features throughout the northeast Pannonian region, the Bohemian Massif, the East Carpathians and the Moesian Platform. Nevertheless, our findings do not emphasize a close correlation between the depth of sedimentary basins and attenuation features identified at longer periods. In addition, Qc variations are larger at short periods, indicating higher heterogeneity in the uppermost crust of Eastern Europe. Our findings demonstrate the higher efficiency of noise correlation approaches relative to earthquake data analyses investigating Qc at low frequencies.

A study of deep neural networks for Newtonian noise subtraction at Terziet in Limburg—the Euregio Meuse-Rhine candidate site for Einstein Telescope

The Euregio Meuse-Rhine border region of Belgium, Germany and the Netherlands has been identified as a candidate site for hosting Einstein Telescope. Newtonian coupling of ground vibrations to the core optics of the detectors may limit the sensitivity of Einstein Telescope at frequencies below about 10 Hz. The contribution of Newtonian noise is site specific and depends on the ambient seismic field which in turn depends on the site’s geology and the distribution of surface and underground seismic-noise sources. We have investigated the application of machine learning in combination with the deployment of seismic sensor networks to predict seismic displacement noise at specific locations on the surface and underground. Moreover we have modeled a deep neural network (DNN) that allows to subtract Newtonian noise from the strain data measured by Einstein Telescope. The seismic-field model is based on a complete solution of the elastodynamic wave equations for a horizontally-layered soil structure. The geology features soft-soil layers on hard-rock and was shown to be effective in attenuating Newtonian noise from surface waves below the required sensitivity. In addition our model includes a random background of body waves with all possible angles of incidence. We show that a DNN is effective in predicting Newtonian noise. Application of our DNN allows Newtonian noise subtraction by a factor up to 4.7 at 1 Hz and 2.5 at 5 Hz.

Near-surface characterization using distributed acoustic sensing in an urban area: Granada, Spain

SUMMARY The Granada Basin in southeast Spain is an area of moderate seismicity. Yet, it hosts some of the highest seismic hazards in the Iberian Peninsula due to the presence of shallow soft sediments amplifying local ground motion. In urban areas, seismic measurements often suffer from sparse instrumentation. An enticing alternative to conventional seismometers is the distributed acoustic sensing (DAS) technology that can convert fibre-optic telecommunication cables into dense arrays of seismic sensors. In this study, we perform a shallow structure analysis using the ambient seismic field interferometry method. We conduct a DAS array field test in the city of Granada on the 26 and 27 August 2020, using a telecommunication fibre. In addition to the existing limitations of using DAS with unknown fibre-ground coupling conditions, the complex geometry of the fibre and limited data recording duration further challenge the extraction of surface-wave information from the ambient seismic field in such an urban environment. Therefore, we develop a processing scheme that incorporates a frequency–wavenumber (f−k) filter to enhance the quality of the virtual shot gathers and related multimode dispersion images. We are able to use this data set to generate several shear-wave velocity (VS) profiles for different sections of the cable. The shallow VS structure shows a good agreement with different geological conditions of soil deposits. This study demonstrates that DAS could provide insights into soil characterization and seismic microzonation in urban areas. In addition, the results contribute to a better understanding of local site response to ground motion.

Temporal variations of the ambient seismic field at the Sardinia candidate site of the Einstein Telescope

SUMMARY Einstein Telescope (ET) is a proposed underground infrastructure in Europe to host future generations of gravitational-wave (GW) detectors. One of its design goals is to extend the observation band of terrestrial GW detectors from currently about 20 Hz down to 3 Hz. The coupling of a detector to its environment becomes stronger at lower frequencies, which makes it important to carefully analyse environmental disturbances at ET candidate sites. Seismic disturbances pose the greatest challenge since there are several important mechanisms for seismic vibrations to produce noise in ET, for example, through gravitational coupling, stray light, or through harmful constraints on the design of ET’s control system. In this paper, we present an analysis of the time-variant properties of the seismic field at the Sardinia candidate site of ET connected to anthropogenic as well as natural phenomena. We find that temporal variations of source distributions and of the noise spectra generally follow predictable trends in the form of diurnal, weekly, or seasonal cycles. Specific seismic sources were identified such as road bridges, which produce observable disturbances underground. This information can be used to adapt a detector’s seismic isolation and control system.

Monitoring the Ambient Seismic Field to Track Groundwater at a Mountain–Front Recharge Zone

The heterogeneity of the fractured-basalt and interbedded-sediment aquifer along the eastern margin of the Columbia Plateau Regional Aquifer System has presented challenges to resource managers in quantifying recharge and estimating sustainable withdrawals. Previous studies indicated recharge pathways in alluvial sediments atop a mountain–front interface upgradient of the basalt flows. In this sedimentary zone, six seismic stations were deployed for one year to detect velocity changes in low-frequency seismic waves that could be correlated to changes in groundwater recorded by a well transducer near the center of the seismic station network. Waveforms in the 1−5 Hz range were recorded at each station to determine changes in wave velocities between station pairs and correlate these velocity changes to changes in groundwater levels. The velocity–groundwater relation allowed for estimation of daily groundwater levels beneath the seismic station network. Existing hydrogeologic information was used to estimate hydraulic gradients and hydraulic conductivities, which allowed for the calculation of the daily volume of recharge passing beneath the seismic stations and into the confined aquifer system. The daily recharge volumes across the seismic station network were summed for comparison of the total annual recharge calculated from the change in seismic wave velocities (154,660 m3) to a flow model calculation of recharge based on areal precipitation and infiltration (26,250 m3). The 6× greater recharge estimated from the seismic wave velocity changes for this portion of the recharge zone is attributed to preferential pathways of high hydraulic conductivity and greater depth associated with paleochannels beneath the seismic station network.

Matched field processing accounting for complex Earth structure: method and review

SUMMARYMatched field processing (MFP) is a technique to locate the source of a recorded wavefield. It is the generalization of plane-wave beamforming, allowing for curved wave fronts. In the standard approach to MFP, simple analytical Green’s functions are used as synthetic wavefields that the recorded wavefields are matched against. We introduce an advancement of MFP by utilizing Green’s functions computed numerically for Earth structure as synthetic wavefields. This allows in principle to incorporate the full complexity of elastic wave propagation without further manual considerations, and through that provide more precise estimates of the recorded wavefield’s origin. We call this approach numerical MFP (nMFP). To demonstrate the applicability and potential of nMFP, we present two real data examples, one for an earthquake in Southern California, and one for secondary microseism activity in the Northeastern Atlantic and Mediterranean Sea. In addition, we explore and clarify connections between localization approaches for the ambient seismic field, real world limitations, and identify key areas for future developments. To increase the adoption of MFP in the seismological community, tutorial code is provided.

Seismic surface wave focal spot imaging: numerical resolution experiments

SUMMARY Numerical experiments of seismic wave propagation in a laterally homogeneous layered medium explore subsurface imaging at subwavelength distances for dense seismic arrays. We choose a time-reversal approach to simulate fundamental mode Rayleigh surface wavefields that are equivalent to the cross-correlation results of three-component ambient seismic field records. We demonstrate that the synthesized 2-D spatial autocorrelation fields in the time domain support local or so-called focal spot imaging. Systematic tests involving clean isotropic surface wavefields but also interfering body wave components and anisotropic incidence assess the accuracy of the phase velocity and dispersion estimates obtained from focal spot properties. The results suggest that data collected within half a wavelength around the origin is usually sufficient to constrain the used Bessel functions models. Generally, the cleaner the surface wavefield the smaller the fitting distances that can be used to accurately estimate the local Rayleigh wave speed. Using models based on isotropic surface wave propagation we find that phase velocity estimates from vertical–radial component data are less biased by P-wave energy compared to estimates obtained from vertical–vertical component data, that even strong anisotropic surface wave incidence yields phase velocity estimates with an accuracy of 1 per cent or better, and that dispersion can be studied in the presence of noise. Estimates using a model to resolve potential medium anisotropy are significantly biased by anisotropic surface wave incidence. The overall accurate results obtained from near-field measurements using isotropic medium assumptions imply that dense array seismic Rayleigh wave focal spot imaging can increase the depth sensitivity compared to ambient noise surface wave tomography. The analogy to elastography focal spot medical imaging implies that a high station density and clean surface wavefields support subwavelength resolution of lateral medium variations.

Recovering and monitoring the thickness, density, and elastic properties of sea ice from seismic noise recorded in Svalbard

Abstract. Due to global warming, the decline in the Arctic sea ice has been accelerating over the last 4 decades, with a rate that was not anticipated by climate models. To improve these models, there is the need to rely on comprehensive field data. Seismic methods are known for their potential to estimate sea-ice thickness and mechanical properties with very good accuracy. However, with the hostile environment and logistical difficulties imposed by the polar regions, seismic studies have remained rare. Due to the rapid technological and methodological progress of the last decade, there has been a recent reconsideration of such approaches. This paper introduces a methodological approach for passive monitoring of both sea-ice thickness and mechanical properties. To demonstrate this concept, we use data from a seismic experiment where an array of 247 geophones was deployed on sea ice in a fjord at Svalbard, between 1 and 24 March 2019. From the continuous recording of the ambient seismic field, the empirical Green function of the seismic waves guided in the ice layer was recovered via the so-called “noise correlation function”. Using specific array processing, the multi-modal dispersion curves of the ice layer were calculated from the noise correlation function, and then inverted for the thickness and elastic properties of the sea ice via Bayesian inference. The evolution of sea-ice properties was monitored for 24 d, and values are consistent with the literature, as well as with measurements made directly in the field.

Newtonian-noise characterization at Terziet in Limburg—the Euregio Meuse–Rhine candidate site for Einstein Telescope

Limburg, in the border region between Belgium, Germany and the Netherlands, has been identified as the Euregio Meuse–Rhine candidate site for Einstein Telescope. The site hosting this gravitational-wave observatory must minimize the Newtonian coupling of ground vibrations to the core optics of the low-frequency detectors. Newtonian noise depends on the ambient seismic field which is in turn dependent on the site’s geology and the distribution of surface and underground seismic-noise sources. We have characterized the site near Terziet in Limburg in terms of propagation modes, dispersion and angular distribution of seismic noise by employing sensor arrays on the surface. Attenuation of seismic noise with depth was studied with a borehole sensor. Based on the results of these measurements, a realistic seismic-field model has been derived that represents a complete solution of the elastodynamic wave equations for a horizontally-layered soil structure. This seismic-field model allows to estimate the Newtonian-noise contribution to the sensitivity of Einstein Telescope for the characteristic geology and ambient noise conditions in South Limburg. The site’s geology features soft-soil layers on hard-rock and is effective in attenuating Newtonian noise from surface waves below the required sensitivity. A random background of body waves with all possible angles of incidence is expected to constitute the dominant source of Newtonian noise.

Towards structural imaging using seismic ambient field correlation artefacts

SUMMARY Correlations of the ambient seismic field recorded by seismic stations carry information about the wave propagation between the stations. They also contain information about the ambient field—both the source of the ambient field, and sources of scattering that contribute to it. The waves that comprise the ambient field are subject to scattering due to the heterogeneous Earth, which can generate supplementary arrivals on the correlation functions. We use these effects to locate sources of signals linked to scattering. For this analysis, we use correlation functions computed from continuous signals recorded between 2013 and 2015 by a line of seismic stations in Central California. We identify spurious arrivals on the Vertical to Vertical and Transverse to Transverse correlation functions and use array analysis to map the source of scattering, which is linked to strong structural variations in the Coast ranges and at the border of the Great Valley.

Solving the century-old mystery of background Love waves

New simulations reveal that much of Earth’s ambient seismic wave field arises from interactions with the planet’s interior.

Accurate Estimations of Sea‐Ice Thickness and Elastic Properties From Seismic Noise Recorded With a Minimal Number of Geophones: From Thin Landfast Ice to Thick Pack Ice

AbstractDespite their high potential for accurate sea ice properties estimation, seismic methods are still limited by the difficulty of access and the challenging logistics of polar environments. Conventional seismic methods generally require tens of geophones together with active seismic sources for monitoring applications. While this is not an issue for mainland environment, it is restrictive for sea ice and prevents long‐term monitoring. We introduce a method to estimate sea ice thickness and elastic properties from passive recordings of the ambient seismic field with a minimal number of geophones. In comparison with our previous work (Moreau et al., 2020; https://doi.org/10.1029/2019JC015709) where about 50 sensors were used, the number of geophones is reduced by 1 order of magnitude, thanks to a new strategy of inversion of the passive seismic data. The method combines noise interferometry for estimating the elastic properties, with a Bayesian inversion of the dispersion in the waveforms of icequakes for inferring ice thickness, based on passive recordings from only 3–5 geophones, depending on the signal to noise ratio. We demonstrate its potential both on data recorded on thin landfast ice in Svalbard, and on data recorded on thick pack ice in the Arctic ocean.

Definition of the Seismic Field of the Underground Sources in the Ambient Seismic Noise in the Tien Shan Region Using a Three-Component Gradient System

This paper considers a new approach to solving the problem of quantitative estimation of the microseism energy for underground sources that is based on the synthesis of noise interferometry and the passive seismic method of the gradient system. The selection of a seismic field of the underground sources is considered in an experiment conducted in the Tien Shan region. The peculiarities of approach include the separation of vertical microseisms in the ambient seismic noise field structure according to the data of the seismic gradient system and a passive noise interferometry diagram, where microseisms from the underground sources are used as the seismic signal source. It is shown that the use of noise interferometry and passive seismic gradient system allows using the synchronous microseism recordings in a small number of points for passive medium sensing, and leads to the restoration of unknown energy parameters of the seismic field of underground sources.

Marine Sediment Characterized by Ocean‐Bottom Fiber‐Optic Seismology

AbstractThe Sanriku ocean‐bottom seismometer system uses an optical fiber cable to guarantee real‐time observations at the seafloor. A dark fiber connected to a Distributed Acoustic Sensing (DAS) interrogator converted the cable in an array of 19,000 seismic sensors. We use these measurements to constrain the velocity structure under a section of the cable. Our analysis relies on 24 hr of ambient seismic field recordings. We obtain a high‐resolution 2‐D shear‐wave velocity profile by inverting multimode dispersion curves extracted from frequency‐wave number analysis. We also produce a reflection image from autocorrelations of ambient seismic field, highlighting strong impedance contrasts at the interface between the sedimentary layers and the basement. In addition, earthquake wavefield analysis and modeling help to further constrain the sediment properties under the cable. Our results show for the first time that ocean‐bottom DAS can produce detailed images of the subsurface, opening new opportunities for cost‐effective ocean‐bottom imaging in the future.

Improving the Retrieval of Offshore‐Onshore Correlation Functions With Machine Learning

AbstractThe retrieval of reliable offshore‐onshore correlation functions is critical to improve our ability to predict long‐period ground motions from megathrust earthquakes. However, localized ambient seismic field sources between offshore and onshore stations can bias correlation functions and generate nonphysical arrivals. We present a two‐step method based on unsupervised learning to improve the quality of correlation functions calculated with the deconvolution technique (e.g., deconvolution functions, DFs). For a DF data set calculated between two stations over a long time period, we first reduce the data set dimensions using the principal component analysis and cluster the features of the low‐dimensional space with a Gaussian mixture model. We then stack the DFs belonging to each cluster together and select the best stacked DF. We apply our technique to DFs calculated every 30 min between an offshore station located on top of the Nankai Trough, Japan, and 78 onshore receivers. Our method removes spurious arrivals and improves the signal‐to‐noise ratio of DFs. Most 30‐min DFs selected by our clustering method are generated during extreme meteorological events such as typhoons. To demonstrate that the DFs obtained with our method contain reliable phases and amplitudes, we use them to simulate the long‐period ground motions from a Mw 5.8 earthquake, which occurred near the offshore station. Results show that the earthquake long‐period ground motions are accurately simulated. Our method can easily be used as an additional processing step when calculating offshore‐onshore DFs and offers a new way to improve the prediction of long‐period ground motions from potential megathrust earthquakes.

Seismic Velocity Changes Caused by Water Table Fluctuation in the New Madrid Seismic Zone and Mississippi Embayment

AbstractWe use cross correlation of the ambient seismic field to estimate seasonal variations of seismic velocity in the Mississippi Embayment and to determine the underlying physical mechanisms. Our main observation is that the δt/t variations correlate primarily with the water table fluctuation, with the largest positive value from May to July and the largest negative value in September/October relative to the annual mean. The δv/v residuals after water level fluctuation correction correlate with air pressure in the short term and follow the trend of temperature in the long term. The δv/v residuals after water level fluctuation and air pressure correction correlate inversely with the wind speed. The correlation coefficients between water table fluctuation and δt/t are independent of the interstation distance and frequency, but high coefficients are observed more often between 0.3 and 1 Hz than between 1 and 2 Hz because high‐frequency coherent signals attenuate faster than low‐frequency ones. The δt/t variations lag behind the water table fluctuation by about 20 days, which suggests that the velocity changes can be attributed to the pore pressure diffusion effect. The seasonal variations of δt/t are azimuthally independent, and a large increase of noise amplitude only introduces a small increase to the δt/t variation. At close distances, the maximum δt/t holds a wide range of values, which is likely related to local structure. At larger distances, velocity variations sample a larger region so that it stabilizes to a more uniform value. We find that the observed changes in wave speed can be explained by a poroelastic model.

Characteristics of the Ambient Seismic Field on a Large-N Seismic Array in the Vienna Basin

Abstract The ambient seismic field is now routinely used for imaging and monitoring purposes. Most commonly, applications aim at resolving crustal-scale features and utilize ocean-generated surface waves. At smaller scales and at frequencies above the microseismic peaks, local sources of seismic energy, often anthropogenic, are dominant, and understanding of their contributions to the ambient seismic field becomes important to apply ambient noise techniques. This study uses data of an industrial-scale seismic deployment covering ∼500 km2 with 10,532 stations, each equipped with several collocated 10 Hz geophones, to provide unique insight into anthropogenic sources of seismic energy in a suburban-to-rural area. We compute amplitude levels, their distance dependency, power spectral densities, and spectrograms to describe the source characteristics. The sources we observe in great detail include windmills, a railway track and trains, cars, oil pumpjacks, power lines, gas pipelines, and airplanes. These sources exhibit time-dependent behavior that is illustrated strikingly by videos of amplitude levels in certain frequency bands that we provide as supplemental material. The data described in this study are a potential resource for future studies, such as automatic signal classification, as well as underground imaging using microseismic noise or the sources presented here.

Ambient seismic field recorded by accelerometers for extracting impulse response functions in the Korean Peninsula

Ambient seismic field has been used intensively to constrain the Earth’s structure since the early 2000s by extracting impulse response functions (IRFs) between two stations. However, mostly IRF phase information was extracted and used rather than the amplitude. Recently, it was expanded to obtain amplitude information as well as phase, particularly by applying the deconvolution method to the ambient seismic field recorded by broadband seismometers. We investigated the potential of ambient seismic field recorded by accelerometers to extract IRFs, including both phase and amplitude information. We analyzed eight stations widely spaced in South Korea, which also contain co-located accelerometer and broadband velocity seismometer, and demonstrate that the ambient seismic field recorded by accelerometers can be successfully used to extract regional surface waves within the period of 2–10 s. We also investigated the period dependent characteristics of extracted IRFs and found the dominant period band of the secondary microseism recorded in the Korean Peninsula. We expect that the ambient seismic field data recorded by accelerometers can provide much denser IRFs on the Korean Peninsula.

NoisePy: A New High-Performance Python Tool for Ambient-Noise Seismology

AbstractThe fast-growing interests in high spatial resolution of seismic imaging and high temporal resolution of seismic monitoring pose great challenges for fast, efficient, and stable data processing in ambient-noise seismology. This coincides with the explosion of available seismic data in the last few years. However, the current computational landscape of ambient seismic field seismology remains highly heterogeneous, with individual researchers building their own homegrown codes. Here, we present NoisePy—a new high-performance python tool designed specifically for large-scale ambient-noise seismology. NoisePy provides most of the processing techniques for the ambient field data and the correlations found in the literature, along with parallel download routines, dispersion analysis, and monitoring functions. NoisePy takes advantage of adaptable seismic data format, a parallel input and output enabled HDF5 data format designed for seismology, for a structured organization of the cross-correlation data. The parallel computing of NoisePy is performed using Message Passing Interface and shows a strong scaling with the number of cores, which is well suited for embarrassingly parallel problems. NoisePy also uses a small memory overhead and stable memory usage. Benchmark comparisons with the latest version of MSNoise demonstrate about four-time improvement in compute time of the cross correlations, which is the slowest step of ambient-noise seismology. NoisePy is suitable for ambient-noise seismology of various data sizes, and it has been tested successfully at handling data of size ranging from a few GBs to several tens of TBs.

Sea Ice Thickness and Elastic Properties From the Analysis of Multimodal Guided Wave Propagation Measured With a Passive Seismic Array

AbstractField data are needed for a better understanding of sea ice decline in the context of climate change. The rapid technological and methodological advances of the last decade have led to a reconsideration of seismic methods in this matter. In particular, passive seismology has filled an important gap by removing the need to use active sources. We present a seismic experiment where an array of 247 geophones was deployed on sea ice, in the Van Mijen fjord near Sveagruva (Svalbard). The array is a mix of 1C and 3C stations with sampling frequencies of 500 and 1000 Hz. They recorded continuously the ambient seismic field in sea ice between 28 February and 26 March 2019. Data also include active acquisitions on 1 and 26 March with a radar antenna, a shaker unit, impulsive sources, and artificial sources of seismic noise. This data set is of unprecedented quality regarding sea ice seismic monitoring, as it also includes thousands of microseismic events recorded each day. By combining passive seismology approaches with specific array processing methods, we demonstrate that the multimodal dispersion curves of sea ice can be calculated without an active source and then used to infer sea ice properties. We calculated an ice thickness, Young's modulus, and Poisson's ratio with values h=54±3 cm, E=3.9±0.15 GPa, and ν=0.34±0.02 on 1 March, and h=58±3 cm, E=4.4±0.15 GPa, and ν=0.32±0.02 on 5 March. These values are consistent with in situ field measurements and observations.