Ambient Seismic Noise Correlations Research Articles

Continuous isolated noise sources induce repeating waves in the coda of ambient noise correlations


 Continuous excitation of isolated noise sources leads to repeating wave arrivals in cross correlations of ambient seismic noise, including throughout their coda. These waves propagate from the isolated sources. We observe this effect on correlation wavefields computed from two years of field data recorded at the Gräfenberg array in Germany and two master stations in Europe. Beamforming the correlation functions in the secondary microseism frequency band reveals repeating waves incoming from distinct directions to the West, which correspond to well-known dominant microseism source locations in the Northeastern Atlantic Ocean. These emerge in addition to the expected acausal and causal correlation wavefield contributions by boundary sources, which are converging onto and diverging from the master station, respectively. Numerical simulations reproduce this observation. We first model a source repeatedly exciting a wavelet, which helps illustrate the fundamental mechanism behind repeated wave generation. Second, we model continuously acting secondary microseism sources and find good agreement with our observations. Our observations and modelling have potentially significant implications for the understanding of correlation wavefields and monitoring of relative velocity changes in particular. Velocity monitoring commonly assumes that only multiply scattered waves, originating from the master station, are present in the coda of the correlation wavefield. We show that repeating waves propagating from isolated noise sources may dominate instead, including the very late coda. Our results imply that in the presence of continously acting noise sources, which we show is the case for ordinary recordings of ocean microseisms, velocity monitoring assuming scattered waves may be adversely affected with regard to measurement technique, spatial resolution, as well as temporal resolution. We further demonstrate that the very late coda of correlation functions contains useful signal, contrary to the common sentiment that it is dominated by instrument noise.

The relationship between cross correlations and Green’s functions in ambient noise interferometry with Bayesian constraints

SUMMARYCross correlation of ambient seismic noise has begun to provide detailed images of the Earth that are not possible to attain using earthquake-based or active-source imaging techniques. However, the theoretical justification for the ambient noise correlation or seismic interferometry approach typically relies on an equipartition or diffuse field assumption that is not expected to be satisfied on Earth. Here we demonstrate that a Bayesian inference approach to calculating the expected cross correlation of seismic signals leads to an improved understanding of its relationship with the Green’s function. The Bayesian derivation is found to exactly replicate the equipartition result under a global energy constraint, and is able to accommodate any number of other constraints that typically cannot be utilized in other approaches. With these stronger constraints, the cross correlation is found to deviate more and more strongly from the ‘expected’ Green’s function, but nonetheless a specific prediction can be made regarding the expected cross correlation which can be used as a more reliable point of comparison with observations than the Green’s function. The approach therefore provides a path forward for how to use seismic cross correlation data under more and more realistic conditions and improvements in knowledge, and should eventually help improve the accuracy of resulting images of the Earth.

Effect of centimetric freezing of the near subsurface on Rayleigh and Love wave velocity in ambient seismic noise correlations

SUMMARY About a decade ago, noise-based monitoring became a key tool in seismology. One of the tools is passive image interferometry (PII), which uses noise correlation functions (NCF) to retrieve seismic velocity variations. Most studies apply PII to vertical components recording oceanic low-frequent ambient noise ( < 1 Hz). In this work, PII is applied to high-frequent urban ambient noise ( > 1 Hz) on three three-component sensors. With environmental sensors inside the subsurface and in the air, we are able to connect observed velocity variations with environmental parameters. Temperatures below 0 °C correlate well with strong shear wave velocity increases. The temperature sensors inside the ground suggest that a frozen layer of less than 5 cm thickness causes apparent velocity increases above 2 % , depending on the channel pair. The observations indicate that the different velocity variation retrieved from the different channel pairs are due to different surface wave responses inherent in the channel pairs. With dispersion curve modelling in a 1-D medium we can verify that surfaces waves of several tens of metres wavelength experience a velocity increase of several percent due to a centimetres thick frozen layer. Moreover, the model verifies that Love waves show larger velocity increases than Rayleigh waves. The findings of this study provide new insights for monitoring with PII. A few days with temperature below 0 °C can already mask other potential targets (e.g. faults or storage sites). Here, we suggest to use vertical components, which is less sensitive to the frozen layer at the surface. If the target is the seasonal freezing, like in permafrost studies, we suggest to use three-component sensors in order to retrieve the Love wave response. This opens the possibility to study other small-scale processes at the shallow subsurface with surface wave responses.

Seismic Noise Correlation on Heterogeneous Supercomputers

ABSTRACT We present a high‐performance tool for the computation of ambient seismic noise correlations on central processing unit (CPU) and graphic processing unit (GPU) clusters. This is intended to address emerging challenges in noise correlation studies with increasingly large data volumes. We propose a parallelization scheme and strategies to efficiently harness modern supercomputing resources, and we demonstrate that the use of GPUs can accelerate the computation of noise correlations by one order of magnitude or more compared with a homogeneous implementation on CPUs. In addition to reducing wall‐clock time, our tool enables on‐the‐fly computations of large noise correlation datasets, thereby eliminating the need for mass storage to archive results.

S-wave velocity model for several regions of the Kamchatka Peninsula from the cross correlations of ambient seismic noise

The data from the seismic networks of the Kamchatka Branch of the Geophysical Survey of the Russian Academy of Sciences are used for calculating the cross correlations of seismic noise for the stationary digital stations over 2013 and for radio telemetric stations (RTS) in the region of the Klyuchevskoy volcano over the period from January 1, 2009 to May 31, 2013. Four hundred and two correlations overall are calculated. The fundamental-mode group velocities of the Rayleigh waves are calculated in the periods ranging from 5 to 50 s. The calculations for the region of the Klyuchevskaya group of volcanoes are based on the RTS data and cover the periods from 2 to 8 s. The two-dimensional (2D) maps of group velocity distributions in different periods are constructed with the use of the algorithm of surface wave tomography (Barmin, 2001). The velocity sections for the selected Kamchatka regions are reconstructed by the dispersion curve inversion technique (Mordret, 2014). For each region, the structure of the Earth’s crust and upper mantle down to a depth of 50 km was obtained.

Along‐strike variations in the Himalayan orogenic wedge structure in Bhutan from ambient seismic noise tomography

AbstractThe geological units and tectonic structure exposed in the Bhutan Himalaya document significant regional variations, expressed primarily as tectonic windows and klippen. The along‐strike variations of these structures and their metamorphic grade are usually associated with the formation of local duplexes in the underlying tectonic units. To investigate these variations and their extent in depth, we image the isotropic shear‐wave velocity structure of the orogenic wedge by ambient noise tomography. Group velocities are extracted from cross correlations of ambient seismic noise, recorded by the temporary GANSSER network in Bhutan. The upper crustal structure beneath Bhutan is mapped down to 18 km depth by directly inverting Rayleigh‐wave group velocity measurements in the period range between 2 and 20 s with a ray tracing based inversion approach. Our results reveal several distinct high shear‐wave velocity anomalies ( km/s) and reflect the along‐strike variations in the upper crustal structure in relation to the alternating tectonic windows and klippen at the surface. In correlation with the surface geology in the northern part of Bhutan, we interpret shallow high shear‐wave velocity anomalies as quarzite‐dominated rocks or felsic migmatites with large intrusions of leucogranites. High‐velocity anomalies in the orogenic wedge in eastern and western Bhutan correlate with the local geometry of the Main Himalayan Thrust and provide evidence for the formation and depth extent of localized duplexes of quartzite dominated lithology in association with the formation of tectonic windows in the Bhutan Himalaya.

Seismic noise monitoring of the water table in a deep-seated, slow-moving landslide

Daily correlations of ambient seismic noise on a large landslide at Utiku, New Zealand, reveal seismic velocity changes up to [Formula: see text] that follow a summer/winter cycle consistent with the pore-water pressures monitored at the basal slip surface in the landslide. The annual pattern of velocity changes is borne by a limited frequency band (6–8 Hz typically) that suggests a localized change in the medium. The Rayleigh waves that form the seismic signal within this frequency band have a maximum sensitivity at a depth of 2–3 m below the ground surface, consistent with the water table level. Fluid saturation changes in the landslide modeled using the Biot-Gassmann theory explain the limited frequency band and the amplitude of the seismic velocity change. This set of arguments suggests that seismic noise correlations are sensitive to water table oscillations through saturation changes and could be used as a nondestructive hydrologic monitoring tool.

Interpreting seismic velocity changes observed with ambient seismic noise correlations

Recent results have shown that crosscorrelating ambient seismic noise recorded in underground mines can successfully extract the seismic Green’s function between sensors. We have revisited an earlier experiment that showed that these virtual seismic sources can be used to measure changes in seismic velocity accurately enough to monitor the short- and long-term influences of a blast in an underground mine. To use this method routinely, it is important to determine the cause of velocity variations in the absence of large dynamic stress perturbations (such as blasts). It also is important to calibrate the seismic velocity changes in terms of known stress changes so the effect of mining activities can be quantified in units that can be used by geotechnical engineers. To this end, we used coda-wave interferometry to measure relative velocity variations during times where no significant blasting or microseismic activity occurred and compared it to atmospheric air pressure changes, temperature variations, and modeled tidal strain. The results indicate that atmospheric air pressure changes have a measurable influence on the long-term seismic velocity variations at depth in the absence of large dynamic stress perturbations. This influence enabled us to determine the sensitivity of the relative velocity changes to stress, where a value of [Formula: see text] was found. This calibration essentially enables us to turn each sensor pair in an underground mine into a stress meter, paving the way for geotechnical engineers to use ambient seismic noise correlations to monitor the evolution of stress and to assess seismic hazard in conjunction with conventional microseismic methods.

Body and surface wave reconstruction from seismic noise correlations between arrays at Piton de la Fournaise volcano

AbstractBody wave reconstruction from ambient seismic noise correlations is an important step toward improving volcano imaging and monitoring. Here we extract body and surface waves that propagate in Piton de la Fournaise volcano on La Réunion island using ambient noise cross correlation and array‐processing techniques. Ambient noise was continuously recorded at three dense arrays, each comprising 49 geophones. To identify and enhance the Green's function from the ambient noise correlation, we apply a double beamforming (DBF) technique between the array pairs. The DBF allows us to separate surface and body waves, direct and reflected waves, and multipathing waves. Based on their azimuths and slownesses, we successfully extract body waves between all the combinations of arrays, including the wave that propagates through the active magmatic system of the volcano. Additionally, we identify the effects of uneven noise source distribution and interpret the surface wave reflections.

Toward 4D Noise‐Based Seismic Probing of Volcanoes: Perspectives from a Large‐NExperiment on Piton de la Fournaise Volcano

Noise‐based seismology is proving to be a complementary approach to active‐source or earthquake‐based methods for imaging and monitoring the Earth’s interior. Until recently, however, noise‐based imaging and monitoring relied mostly on the inversion of surface waves reconstructed from correlations of mostly microseismic seismic noise (around 5 s of period). Compared to body‐wave tomography and considering similar wavelengths, surface‐wave tomography succeeds in retrieving lateral subsurface velocity contrasts but is less efficient in resolving velocity perturbations at depth. Recently, the use of large‐ N seismic arrays has proven to be of great benefit for extracting body waves from noise correlations by stacking over a large number of receiver pairs and by applying array processing. In this article, we describe a recent large‐ N array experiment that we conducted on Piton de la Fournaise (PdF). Our main goal was to extract body waves traveling directly in the vicinity of the active magma reservoir located at ∼2.5 km depth below the summit crater using noise correlations between arrays of seismic sensors. Within this article, we provide technical information about the VolcArray experiment, which consisted of the deployment of 300 seismic nodes during one month on PdF. We also present data‐quality measurements and show how the short‐period seismometers that we used compare to standard broadband seismic records. Finally, we show noise array beamforming results to study the content of seismic noise at frequencies between 1 and 12 Hz and discuss the ability to use these data to recover body waves between arrays from the correlation of ambient seismic noise.

Investigation of coseismic and postseismic processes using in situ measurements of seismic velocity variations in an underground mine

AbstractThe in situ mechanical response of a rock mass to a sudden dynamic and static stress change is still poorly known. To tackle this question, we conducted an experiment in an underground mine to examine (1) the influence of dynamic and static stress perturbations on seismic velocities, (2) elastic static stress changes, and (3) induced earthquake activity associated with the blast and removal of a portion of hard rock. We accurately (0.01%) measured seismic velocity variations with ambient seismic noise correlations, located aftershock activity, and performed elastic static stress modeling. Overall, we observe that the blast induced a sudden decrease in seismic velocities over the entire studied area, which we interpreted as the damage due to the passing of strong seismic waves. This sudden process is followed by a slow relaxation lasting up to 5 days, while seismic activity returns to its background level after 2 days. In some locations, after the short‐term effects of the blast have subsided, the seismic velocities converge to new baseline levels and permanent changes in seismic velocity become visible. After comparing the spatial pattern of permanent seismic velocity changes with elastic static stress modeling, we infer that the permanent seismic velocity changes are due to the change in the static volumetric stress induced by the removal of a solid portion of rock by the blast. To our knowledge, this is the first observation of noise‐based permanent seismic velocity changes associated with static stress changes.

Body-wave reconstruction from ambient seismic noise correlations in an underground mine

The reconstruction of seismic Green’s functions from correlations of ambient seismic noise has recently developed as a promising approach for exploring the earth’s interiors without the requirement of costly active seismic sources. This approach is widely used for imaging the crust at a kilometer scale. However, few studies report noise-based Green’s function reconstruction at smaller scales in industrial environments. We have investigated the possibility of constructing seismic Green’s functions between sensors in an active underground mine (Garpenberg, Sweden) by crosscorrelating seismic noise generated by mining activities. We have determined with realistic numerical simulations that the mining excavations in an underground mine lead to a regime of strong scattering, which is favorable for constructing seismic Green’s functions by crosscorrelating seismic noise. One month of continuous data was recorded by 18 seismic sensors located more than 1 km below surface. A variety of broadband (10–3000 Hz) seismic sources were present, but the seismic wavefields are directional and often monochromatic, so that the conditions for constructing Green’s functions by crosscorrelating ambient seismic noise were not ideal (isotropic illumination and spectrally white). We developed a stacking scheme that dismissed data during periods when the seismic noise was dominated by monochromatic signals or when noise sources were not in stationary phase locations. Estimates of the seismic Green’s functions were retrieved for a broad frequency range (20–400 Hz) for almost all of the correlation pairs when we used the selective stacking scheme. We used the direct body waves present at low frequencies (less than 100 Hz) in the reconstructed seismic Green’s functions to invert for the 3D S-wave velocity structure of the mine. Our results revealed the existence of a high-velocity zone and a low-velocity zone that corresponded with known ore bodies.

Ambient seismic, hydroacoustic, and flexural gravity wave noise on a tabular iceberg

AbstractCross correlation of ambient seismic noise between four seismographs on tabular iceberg C16, Ross Sea, Antarctica, reveals both the source and the propagation characteristics of signals associated with icebergs. We find that noise correlation functions computed from station data are asymmetric about zero time lag, and this indicates that noise observed on the iceberg originates primarily from a compact, localized source associated with iceberg collisions between C16 and a neighboring iceberg, B15A. We additionally find two, and possibly more, distinct phases of noise propagation. We believe that flexural gravity wave propagation dominates the low‐frequency noise (>10 s period) and that hydroacoustic wave propagation in the water column between the ice and seabed appears to dominate high‐frequency noise (>10 Hz). Faster seismic propagation dominates the intermediate band (2–6 Hz); however, we do not have sufficient data to characterize the wave mechanisms more precisely, e.g., by identifying distinct longitudinal and shear body waves and/or surface waves. Secular changes in the amplitude and timing of ambient noise correlations, e.g., a diurnal cycle and an apparent shift in the noise correlation of fast seismic modes between two periods of the deployment, allow us to speculate that ambient noise correlation analysis may be helpful in understanding the sources and environmental controls on iceberg‐generated ocean noise as well as geometric properties (such as water column thickness) of subglacial lakes.

Long‐term dynamics of Piton de la Fournaise volcano from 13 years of seismic velocity change measurements and GPS observations

AbstractWe study the Piton de la Fournaise (PdF) volcano dynamics through the observation of continuous seismic velocity changes from 2000 to 2013. We compute the cross correlations of ambient seismic noise recorded at more than 30 short‐period and broadband stations of the UnderVolc temporary seismic experiment and of the PdF volcano observatory network. The velocity changes are estimated from the travel time delay measured on the cross correlations computed between pairs of stations. We average the relative velocity changes for all pairs of stations and obtain a time series of the velocity change of Piton de la Fournaise volcano over 13 years. From the period 0.5 to 4 s, the depth sensitivity of the velocity change is ranging from approximately 100 m to 2500 m. A slow decrease of velocity is measured from 2000 and ends with a major eruption that occurred in April 2007. This eruptive episode is followed by an increase of the velocity. These long‐term changes are compared to the deformation of the Piton de la Fournaise edifice estimated from geodetic measurements. An analysis of baseline change between GPS stations indicates an inflation of the volcanic edifice prior to April 2007 followed by a deflation since then. This deflation predominantly affects the terminal cone. Seismic velocity changes and deformation have similar long‐term trends with velocity decrease observed during inflation and velocity increase during deflation. However, the velocity change magnitude is about 2 orders of magnitude greater than the deformation. This suggests nonlinear relation between velocity changes and deformation.

Correction of Ocean-Bottom Seismometer Instrumental Clock Errors Using Ambient Seismic Noise

Abstract Very accurate timing of seismic recordings is critical for modern processing techniques. Clock synchronization among the instruments constituting an array is, however, difficult without direct communication between them. Synchronization to Global Positioning System (GPS) time is one option for on‐land deployments, but not for underwater surveys as electromagnetic signals do not propagate efficiently in water. If clock drift is linear, time corrections for ocean‐bottom seismometer (OBS) deployments can be estimated through GPS synchronization before and after the deployment, but this is not sufficient for many applications as the nonlinear component of the drift can reach tens to hundreds of milliseconds for long‐duration experiments. We present two techniques to retrieve timing differences between simultaneous recordings at ocean‐bottom instruments after deployment has ended. Both techniques are based on the analysis of the cross correlation of ambient seismic noise and are effective even if clock drift is nonlinear. The first, called time symmetry analysis, is easy to apply but requires a proper illumination so that the noise cross‐correlation functions are symmetric in time. The second is based on the doublet analysis method and does not have this restriction. Advantages and drawbacks of both approaches are discussed. Application to two OBS data sets shows that both can achieve synchronization of recordings down to about five milliseconds (a few percent of the main period used).

Teleseismic correlations of ambient seismic noise for deep global imaging of the Earth

Abstract We present here a global analysis showing that wave paths probing the deepest part of the Earth can be obtained from ambient noise records. Correlations of seismic noise recorded at sensors located various distances apart provide new virtual seismograms for paths that are not present in earthquake data. The main arrivals already known for earthquake data are also present in teleseismic correlations sections, including waves that have propagated through the Earth's core. We present examples of applications of such teleseismic correlations to lithospheric imaging, study of the core mantle boundary or of the anisotropy of the inner core.

Body-Wave Imaging of Earth’s Mantle Discontinuities from Ambient Seismic Noise

Ambient seismic noise correlations are widely used for high-resolution surface-wave imaging of Earth's lithosphere. Similar observations of the seismic body waves that propagate through the interior of Earth would provide a window into the deep Earth. We report the observation of the mantle transition zone through noise correlations of P waves as they are reflected by the discontinuities associated with the top [410 kliometers (km)] and the bottom (660 km) of this zone. Our data demonstrate that high-resolution mapping of the mantle transition zone is possible without using earthquake sources.

Understanding the amplitudes of noise correlation measurements

Cross correlation of ambient seismic noise is known to result in time series from which station-station travel-time measurements can be made. Part of the reason that these cross-correlation travel-time measurements are reliable is that there exists a theoretical framework that quantifies how these travel times depend on the features of the ambient noise. However, corresponding theoretical results do not currently exist to describe how the amplitudes of the cross correlation depend on such features. For example, currently it is not possible to take a given distribution of noise sources and calculate the cross correlation amplitudes one would expect from such a distribution. Here, we provide a ray-theoretical framework for calculating cross correlations. This framework differs from previous work in that it explicitly accounts for attenuation as well as the spatial distribution of sources and therefore can address the issue of quantifying amplitudes in noise correlation measurements. After introducing the general framework, we apply it to two specific problems. First, we show that we can quantify the amplitudes of coherency measurements, and find that the decay of coherency with station-station spacing depends crucially on the distribution of noise sources. We suggest that researchers interested in performing attenuation measurements from noise coherency should first determine how the dominant sources of noise are distributed. Second, we show that we can quantify the signal-to-noise ratio of noise correlations more precisely than previous work, and that these signal-to-noise ratios can be estimated for given situations prior to the deployment of seismometers. It is expected that there are applications of the theoretical framework beyond the two specific cases considered, but these applications await future work.

Reconstructing the Green's function through iteration of correlations

Correlations of ambient seismic noise are now widely used to retrieve the Earth response between two points. In this study, we reconstruct the surface-wave Green's function by iterating the correlation process over the tail of the noise-based correlation function. It has been demonstrated that the so-called C 3 function shows the surface-wave part of the Green's function. Using data from 150 continuously recording stations in Europe, the C 3 results help in the extraction of the travel-times from noise-based measurements, especially through the suppression of effects caused by non-isotropic source distributions. We present the results of the next iterative step (i.e. C 5), which show that some coherent signal is still present in the coda of the C 3 function, and we investigate the evolution of the reconstruction of the Green's function throughout the iteration process. Finally, we discuss the interest of combining information from the different correlation functions to improve noise-based tomography analysis.

Shallow structure of Deception Island, Antarctica, from correlations of ambient seismic noise on a set of dense seismic arrays

SUMMARY We investigated the shallow velocity structure of Deception Island volcano, Antarctica, using correlations of ambient seismic noise. We selected long records of noise obtained by eight seismic arrays deployed along the inner coast of Deception during the period 2003–2005. Using these data, we calculated average dispersion curves and estimated local 1-D velocity models for the array sites. The combination of these profiles allowed us to obtain a comprehensive model of the shallow velocity structure of the island. The volcano is composed of relatively soft layers of pyroclastic deposits and sediments extending to a depth of about 400 m, with different degrees of compaction. Two layers with thicknesses of about 100 and 300 m and S-wave velocities of around 0.2–0.8 and 0.7–1.1 km s−1, respectively, can be differentiated. The deeper structure is highly variable in terms of wave velocities and layer depths. Although the resolving capabilities are reduced for these layers, the larger S-wave velocities in the range 1.3–2.8 km s−1 indicate that they can be associated with pre-caldera structures and products. There are substantial differences between the different models, which can be spatially related to heterogeneities in the volcano structure. The lowest S-wave velocities may be related to the alterations produced by hydrothermal activity near the surface. On the contrary, the largest velocities occur near the caldera border, revealing the presence of compact materials at shallow depths. Sharp lateral variations can also be observed in the northwest of the bay, which points to the presence of NW–SE faults and/or strong velocity gradients.