Posters

Walter Fabian

SEISMIC NOISE ON THE GREENLAND ICE SHEET

Over the last decade, the interest in glacier seismology has drastically increased. To date, seismic monitoring has provided valuable information about subglacial processes, iceberg calving and ice deformation. Seismic investigations are an attractive alternative to traditional in-situ glaciological measurements. So-called ‘icequakes’ reveal details about englacial fracturing, basal processes and englacial water flow. These events can be located with high-density seismic networks, which allows for monitoring of ice flow processes over an extended glacier region. Most glacier-related seismic investigations have focused isolated seismic transients emitted by discrete seismic events. However, recent investigations have highlighted continuous seismic sources within glacier ice with characteristics of sustained seismic tremor that resemble those seen in volcanic environments. Accordingly, icequake location and waveform modeling techniques cannot be applied to these signals. On the other hand, glacial tremor location and signal characteristics are of supreme importance, as these parameters elucidate englacial water flow and changes thereof. Such water flow can perturb subglacial conditions and hence critically alter ice flow. Here we apply linear and adaptive beam-forming techniques to continuous seismic noise records from the Greenland ice sheet. The signals were recorded with a high-density campaign seismic network in summer 2011 on the ablation zone near Jakobshavn Ice Stream. The noise records contain highly coherent signals, whose origin likely depends on the chosen frequency band. We discuss these findings in the context of melt water drainage and transient changes in glacier dynamics.

Olivier Gerrit

Monitoring in mines with ambient seismic noise

In crustal studies the ambient seismic noise is dominated by surface waves emanating from the interaction of the ocean with the solid earth. In an underground mining environment this is not the case, since the strength of these surface waves decreases rapidly with depth. The noise is instead dominated by human activity associated with mining activities, which is generally rather localised, short lived and spatially unstable. In this presentation I discuss the possibility of using ambient noise generated by mining activities to monitor relative travel time variations between sensors. The results from a mine scale numerical experiment is shown, in which a finite-difference dynamic seismic wavefield code was used to generate synthetic seismograms in the presence of realistic mine tunnels and excavations. Actual noise from drilling activities was used as a source in the numerical experiment. The results confirm that the ambient noise method should be useful in a typical underground mine. We also investigate the effects of noise source stability on the accuracy of the measurements and discuss how a small sensor array could be used with beamforming to determine the direction of the dominant noise source. Three weeks of continuous seismic data from 4 sensors in an underground mine 1 km below surface has been analysed. We find that 3 minute segments of ambient noise in the frequency band between 5 and 400 Hz give surprisingly stable cross-correlation functions. This enables us to make robust regular estimates of relative travel time variation with a accuracy as small as 0.001 % (10^-5). The origins for such small temporal travel time perturbations are explored by comparing these changes with barometric pressure, temperature, sea level temporal changes and tide-induced deformation.
Cambaz Didem
SURFACE WAVE TOMOGRAPHY OF TURKEY AND SURROUNDINGS FROM AMBIENT SEISMIC NOISE CORRELATION
We perform ambient noise tomography of Turkey using data from the National Seismic Network of Turkey, temporary stations and surrounding regional stations. For most of the station pairs, we retrieve good Rayleigh waveforms from ambient noise correlations using 2-6 years of continuous data at all distance ranges across the entire region (>1500 km) for periods from 10s to 40s. We obtain Rayleigh wave group velocity dispersion measurements using a frequency-time analysis and obtain Rayleigh wave group velocity maps for periods from 10 s to 40 s. The maps indicate the presence of significantly different crustal compositions and structures resulting from different tectonic evolutions. Group velocity maps exhibit strong velocity perturbations and correlates well with the known tectonic structures. In general, the tomographic images at short periods (10-15s) displaying low velocities associate with the sedimentary basins, intermediate periods (20-30sec) with regional geologic structures and greater periods (40-50) total crustal structure and upper mantle.

Brenguier Florent

Study of the Great 2011 Tohoku earthquake using ambient seismic noise

We explore crustal seismic velocity temporal changes in Japan associated with the great 2011 Tohoku earthquake using ambient seismic noise.

Korakas Alexios

Matched-field inversion in a three-dimensional wedge-like environment

Inverse problems in ocean acoustics are based on two-dimensional (2-D) modeling of sound propagation, hence ignoring the effects of horizontal refraction, referred to as three-dimensional (3-D) propagation effects. However, the acoustic propagation in shallow-water environments, such as the continental shelf, may be affected by 3-D effects requiring 3-D modeling to be accounted for. At first glance, the inversion of such propagation data would suggest the use of a fully 3-D model, making the inversion impractical for more than two parameters at a time due to the dramatically increased CPU times. In the present poster, it is shown that a 2-D-model-based inversion of simulated vertical array data of 3-D propagation over a sloping bottom, succeeds in retrieving the correct parameter values at relatively short ranges from the source. Applying this observation to the inversion of laboratory scale measurements of long range propagation in a 3-D wedge leads to a subspace inversion approach that yields excellent comparisons with solutions by a fully 3-D parabolic equation-based model in reasonable CPU times.

Macquet Marie

Ambiant noise tomography of the crustal structure beneath the Pyrenees

We use data of a dense temporary seismic broadband array of 49 stations installed in southwestern France and along the Atlantic coast at the end of 2010 in the framework of the PYROPE (PYRenean Observational Portable Experiment) project. We also used records of 70 broadband stations of the third leg of the temporary IberArray project installed in northern Spain. To complement the two temporary arrays and avoid smearing effects along their edges, our dataset also includes records of the French and Catalan permanent broadband networks. As a whole, the dataset includes records of 158 broadband stations. Using one year of data, we computed 12324 two-station correlations of ambient seismic noise records. We then measured group velocity dispersion of the fundamental mode of the Rayleigh wave in the period range 5 to 30 s. These group velocity curves are subsequently used to obtain 2-D group velocity maps. The last step of our tomography is the inversion of local group velocity dispersion curves for the 3-D Vs structure. The resulting velocity model will additionally to geodynamic purpuses improve the precision of hypocentral locations in the region.

Mangongolo Azangi

Toward ambient noise studies of the Water Ingress, South Africa

Mordret Aurelien

On the use of ambient noise surface-wave tomography in marine exploration geophysics context

We present Scholte wave phase velocity maps obtained from cross-correlation of 6.5 hours of noise data from the Valhall Life of Field Seismic (LoFS) network. We computed 2 690 040 vertical-vertical component cross-correlations from the 2320 available sensors, turning each sensor into a virtual source emitting Scholte waves. For every virtual source, we measured the Scholte wave travel times to all other stations and interpolated them on a regular grid to perform a ‘Helmoltz tomography’ of the Valhall oil field subsurface. This tomography technique takes advantage of very dense seismic network like Valhall LoFS network without the need of a formal inversion and automatically accounts for bent rays and finite frequency effects which makes it more accurate comparing with standard straight-ray surface wave tomography methods. Scholte wave isotropic phase velocity maps at periods between 0.65 s and 1.6 s show a coherent geomorphological pattern dominated by paleo-channels in the shallower part. In the deeper part we observe a hint of gas infiltration as a very low velocity anomaly. The inversion of these maps in depth allowed us to access the S-wave velocity distribution in the first hundreds of meters of the Valhall subsurface. We also retrieved the azimuthal anisotropy and its lateral variations showing a characteristic elliptical pattern around the central exploitation platform. These results would have important implication for shear wave statics and monitoring of subsidence due to oil extraction.

Geli Louis

Monitoring of gas emissions in deep sea environments using OBSs and active acoustic gas bubble recorder

Frank William

Low-frequency earthquakes in the Mexican Sweet Spot

We use data from the MesoAmerican Subduction Exper- iment that operated between 2005 and mid-2007 to detect and locate low-frequency earthquakes (LFEs) in the Guer- rero region in Mexico. We start by identifying a set of visi- ble events and use them as templates to perform a network waveform correlation search of the continuous seismic data. This procedure results in ∼17,000 robustly detected LFEs regrouped into 15 distinct families based on waveform simi- larity. First, we study the evolution of the LFE activity in time and observe a nearly constant background LFE activity with intensifications during the large 2006 slow-slip event as well as previously reported periods of increased non-volcanic tremor (NVT) activity. We stack all robust detections cor- responding to the same LFE family. This procedure results in seismograms with high signal-to-noise ratios that can be used to pick the arrival times of the P- and S-waves. We use these travel times to locate the LFE sources by com- paring their relative moveouts to the moveouts computed for a grid of potential sources. The resulting locations su- perpose a previously identified region of permanent NVT activity (reported as a Sweet Spot by Husker et al. [2012]) indicating a strong link between the two phenomena. All of the LFE hypocenters have been located at a depth of 40 - 45 km in an area that is surrounding the upper slab-plate interface that has previously been suggested to be a region of higher pore-pressure that facilitates NVT generation. We then finally characterized their focal mechanisms by compar- ing their stacked seismograms to synthetic seismograms with varying moment tensors. A common low-dipping thrusting focal mechanism was identified that reflects the first-order Mexican subduction regime.

Lontsi Agostiny Marrios

Combined inversion of seismic ambient vibrations phase velocity and in depth H/V(f,z) spectral ratio for site characterization

A seismic ambient noise survey was carried out at the TTS (Technical Security Test site) of the BAM (Federal Institute for Materials Research and Testing) in Horstwalde, around 50 km south of Berlin within a glaciofluvial sedimentary environment to delineate the sediment thickness and the 1D shear-wave velocity profile. In addition to three permanent 3C-boreholes available at the site, we have used a non-redundant linear array configuration of six Lennartz LE-3D 1s sensors recording at 200 Hz sampling rate over a maximum aperture of 1760 m. Following the principle ideas of seismic interferometry we computed long-term average cross-correlation for all 15 resulting receiver-pairs along the line. Assuming the equivalence of the cross-correlations with inter-station Green’s functions we can create a virtual linear co-array configuration with approx. 100 m inter-station spacing that allows processing the data similar to active MASW ("multi-channel analysis of surface waves") recordings. Two modes are identified with exploitable frequency range and sufficiently high signal to noise ratio between 0.3 Hz to 4Hz and 0.8 Hz to 1.7Hz for the fundamental and first higher mode respectively. In contrast to an active recording experiment, the frequency content of the ambient vibration wave-field is enriched in low frequencies thus providing better resolution for deeper structures. Under the assumption of a mainly horizontally layered 1D-structure we further use the result of single station H/V analysis at each recording instrument at different depths together with the dispersion curves to invert for the S-wave velocity profile using the direct search neighbourhood algorithm for the parameter space sampling. We compare the result of the shallow part to those obtained from shallow active seismic experiments and direct push measurements that are available to us for the study region. The overall thickness is compared to that deduced from P-SV converted delay time using receiver function analysis of mine induced Earthquake recorded at a broadband seismic station at the test site and also to that from empirical relation available from the literature.

Obermann Anne

Locating changes on an active volcano using ambient seismic noise cross-correlations

We analyze continuous ambient seismic noise records from the active volcano Piton de la Fournaise on la Reunion Island from June to December 2010. During this time two volcanic eruptions occurred. We calculate the cross-correlation functions between 21 broadband stations for this period. We monitor the apparent relative velocity changes with the stretching technique, as well as the evolution of the decoherence of the signals. The temporal variations of both parameters are precursors of volcanic eruptions. We compute sensitivity kernels between the stations using a radiative transfer approach for the intensity propagation. Then we use a least-square inversion to locate the changes associated to the eruptions (apparent velocity changes and waveform decorrelation) in space : prior, during, and after the eruption.

Daub Eric

Ambient noise imaging and nonlinear elasticity of the crust during faulting

Ambient seismic noise provides a powerful tool for monitoring temporal changes in the mechanical properties of the crust. Studies utilizing ambient seismic noise have been conducted in a number of settings, providing data on time-dependent changes in seismic wave speeds during a variety of faulting events. While these observations show that small changes in the mechanical properties of the crust can be reliably measured using noise, it is not a clear how to interpret these changes and how to relate these changes to the stresses and strains of faulting events. Here, we use ambient seismic noise to explore time-dependent changes of seismic wave speeds during various faulting events, including earthquakes and aseismic slow slip events. We relate the changes to observed strains and strain rates, as well as the time-dependent recovery following faulting, and encapsulate our results in a nonlinear elastic model framework. Through this model, we can compare to analogous laboratory measurements on the nonlinear elastic properties of rocks and other materials. Our results show that the crust exhibits a complex, nonlinear elastic behavior, and provides useful constraints for how to interpret monitoring studies of faults and other systems.

Colombi Andrea

The complete Green’s function of a complex medium measured by interferometry : reverberated flexural waves on a thin aluminium plate

Using flexural waves on a thin plate we propose an analysis supported by laboratory and numerical experiments that aim to disentangle the relative roles of scattering, reverberation and source distribution in giving rise to a wave field diffuse enough for cross-correlations of ambient signal to approximate well the Green’s function.

Retailleau Lise

Focusing of the seismic waves at the epicenter’s antipode observed with the USArray

We present an analysis of records by the USArray stations of an earthquake that occurred on the Southeast Indian Ridge on February 11, 2010. The epicenter of this event is antipodal to the USArray providing us with an opportunity to observe in details the antipodal focusing of seismic wave in space and time. We compare the observed signals with synthetic seismograms computed in different frequency bands for a spherically symmetric Earth model (PREM). Deviation of the observed focusing patterns from those predicted with synthetic seismograms is a direct evidence of the lateral heterogeneity within the Earth. The behavior of strong P-wave phases (PKP, PP, PPP) is well predicted by the spherically symmetric model. We start to observe significant deviations for converted phases (PPS, SKSP) and the focusing deteriorates very strongly for S waves propagating in the mantle.

Hanson-Hedgecock Sara

Measuring velocity changes associated with increased eruptive activity at Tungurahua Volcano, Ecuador in 2010, from ambient noise correlations

We present the results of measuring velocity changes associated with three episodes of increased eruptive activity at Tungurahua in 2010 using ambient noise correlations. Tungurahua volcano is one of the most active volcanoes in the Ecuadorian Andes. Eruptions since 1999 have generated Strombolian and Vulcanian style volcanic activity that pose a threat to 25,000 people livings in the towns (including Ba√±os) that surround the volcano, as well as the Agoyan hydroelectric dam. During 2010 Tungurahua experienced three episodes of heightened eruptive activity in (1) January to early March, (2) Late May to June 18th paroxysmal crisis with continued moderate activity through July, and (3) mid-November to December. Following standard procedures, ambient noise correlations are calculated from daily, 0.5-4Hz bandpass filtered, continuous seismic records at 5 broad-band seismometers, deployed around the flanks of Tungurahua by the Instituto Geof√≠sico Escuela Polit√©cnica Nacional (IGEPN). Using the dilation correlation coefficient (stretching) method [Lobkis and Weaver, 2003, Hadziioannou et al., 2009, 2011, Weaver et al. 2011] the homogenous relative velocity change, dV/V , is calculated that maximizes the cross-correlation coefficient between each daily correlation function and a reference function composed of the 1 year-long stack of the daily correlations. During the three eruptive episodes of 2010, a drop in the relative velocity is observed of 0.7%, 0.5%, and 0.4%, respectively. These values are greater than the calculated root mean square of the erroneous relative dilation, rms Œµ = 0.2%. The observed drop in relative velocity begins at the onset of eruption, decreases gradually, and begins to recover as eruptive activity decreases. The coinciding of decreased velocity with increased eruptive activity and number of recorded LPs earthquakes suggests a source that is directly related to eruption processes, e.g., fracturing and dilation due to pressurization in the magmatic/hydrothermal system, shallowing of fragmentation depth, or deepening of the exsolution depth.

Landes Matthieu

Studying seismic noise sources generating shear waves

The ambient seismic noise is dominated by the secondary microseism peak (5-7s) generated by the non linear interaction of oceanic waves.With the development of numerical wave action models, the modeling of microseism allows to explain different observations as the shape and amplitude of seismic spectrum, source location and signals generated by specific events like storms. However, these studies concentrate on compressional waves while the observations of shear waves could raise new questions : Love waves are part of the ambient seismic noise, but what about S waves ? Is the compressional and shear waves generated systematically by the same mechanism and with the same source distributions ? Should we distinguished body and surface waves ? In this study, we provide observations that compares beamforming results on vertical and horizontal components. Data are continuous noise records between January 2007 and July 2008 of an Indian network. The comparison of the azimuthal distribution of surface wave observed at the network between Rayleigh and Loves waves and the comparison of source location of P and S waves highlight two different behaviours for compressional and shear wave sources that generates the background seismic noise. These observations suggest issues for the source mechanism at the Earth/Ocean interface.

Boue Pierre

Probing deep Earth with body waves from ambient seismic noise correlation

Seismic noise correlation is now an established method to observe the propagation of surface waves between pairs of sensors, without involving transient sources. At low frequency, these observed surface waves are exploited to depict high-resolution image of the crust and upper mantle, or mapping the velocity changes associated with tectonic events. Recent works highlight more challenging observations such as attenuation measurement and body waves retrieval at regional scale. Here we focus on the detection of body waves at teleseismic distance using a global broadband dataset. We first show that body waves emerge from cross-correlations for every pairs of seismic stations whatever their geographical locations. Such worldwide body wave observations provide new and original information on the Earth structure and could now be used to imaging purposes, by applying classical methods for deep Earth study. In this work, we compare the global time-vs-offset representation of the cross-correlations computed in different period bands between 1 and 150s for which different phases are preferentially reconstructed. Then we analyze and discuss the deep Earth phases such as core phases which are usually extracted and studied from the earthquake seismograms.

Rivet Diane

Insights on the deep activity of Piton de la Fournaise Volcano from long-term seismic velocity changes

We study Piton de la Fournaise (PdF) Volcano dynamics through the observation of continuous seismic velocity changes during the period 2009 to 2013. We compute cross-correlations of ambient seismic noise recorded at 21 broad-band stations of the Undervolc and PdF Volcano Observatory networks. The velocity changes are estimated from the travel time delay measured on the coda of noise cross-correlations. The portion of the coda we use to measure velocity changes consists predominantly of surface waves. The sensitivity at depth of the coda waves thus varies with the frequency, similarly as the one of the Rayleigh waves. The short period waves are sensitive to the shallow structure of the volcano, while the long period waves are sensitive to its deeper structure. Using this property of the surface waves, we seek to estimate the velocity perturbation at different depth. We observe a dependence of seismic velocity change with the period band considered. At short period (0.5-4s), in addition to the short-term velocity changes produced by the volcanic eruptions (October 2010 and December 2010), a long-term increase of velocity is measured between 2009 and 2013. This is consistent with geodetic measurements, which indicate a deflation of the volcanic edifice since April 2007. At longer periods we observe velocity changes that do not correlate in time with PdF eruptive activity. Yet, distinctive episode of velocity changes are observed at different period bands suggesting changes in the elastic properties at different depth.

Hobiger Manuel

Depth distribution of coseismic velocity changes caused by large earthquakes in Japan

Using Passive Image Interferometry, i.e. by cross-correlating ambient seismic noise recorded by Hi-net sensors, we measured coseismic and postseismic velocity changes for several earthquakes (Mw > 6.5) which occurred in Japan (Fukuoka 2005, Noto Hanto 2007, Iwate-Miyagi Nairiku 2008, three earthquakes in Niigata prefecture 2004, 2007 and 2011). For each of these earthquakes, we cross-correlated ambient seismic noise recordings of several years in four different frequency ranges between 0.125 and 2.0 Hz. For each earthquake, the data of 10 to 20 seismometers with distances of up to 50 km from the epicenter have been analyzed, forming between 45 and 190 station pairs. Using a simple tomography algorithm, the observations of the station pairs can be reprojected onto the actual station locations. For the analyzed earthquakes, the observed coseismic velocity changes are systematically larger at higher frequency. As the analyzed seismic noise is mainly composed of surface waves, these findings suggest that the coseismic velocity changes are concentrated in the shallow layers of the ground structure. Using the coseismic velocity changes at different frequencies, the actual depth distribution of the coseismic velocity changes can be modeled. Starting from a reference shear and pressure wave velocity profile, the corresponding Rayleigh wave dispersion curve is modified according to the measured velocity changes in the different frequency ranges. By changing the original velocity profile in a simple way (constant percental velocity change between given depths), the depth distribution of the coseismic velocity changes can be constrained. For the Iwate-Miyagi Nairiku earthquake, for example, the coseismic velocity change is most likely to be of the order of -3% and to be concentrated in the shallowest several hundred meters.

Cros Estelle

Improving the accuracy of earthquakes localization using OBSs in the Sea of Marmara, Turkey

The Marmara Sea is located between the Agean Sea and the Black Sea, along the North Anatolian strike-slip fault, which experienced a sixty year sequence of earthquakes since 1940. Prior to this sequence, which ended with the Izmit and Duzce earthquakes in 1999, at the eastern end of the Sea of Marmara (SoM), the fault ruptured to the west in 1912 in Ganos, with an estimated moment magnitude of 7.4. Therefore, a major earthquake is expected within the SoM seismic gap. In order to better understand the seismicity and to reduce the threshold of detection, a network of ten OBS with four components was deployed by Ifremer with R/V Yunus of Istanbul Technical University, in the western and central parts of the Marmara Sea to record the micro-seismicity from the immediate vicinity of the main Marmara Fault, between April and August, 2011. The network was specifically designed to survey the segments crossing the Western High, where gas hydrates where recently found, the Central Basin and the Kumburgaz Basin. During this period more than one hundred earthquakes were detected by the EMSC (European-Mediterranean Seismological Centre) in the Sea of Marmara. Because the basins of the Sea of Marmara are filled with more than 5 km of Plio- Quaternary soft (“slow”) sediments, it is of critical importance to take into account the velocity structure of the offshore domain, which is drastically different from the one onshore, and the bathymetry. To improve the localization of seismic events, a 3D velocity model was thus considered and implemented in the Sytmis software developed by INERIS. This model is based on the tomographic data collected in 2001 using a controlled source experiment and on the numerous multichannel seismic profiles that provide information on, respectively, the deeper structures and the upper, sedimentary layers. Preliminary results are presented. Special focus will be given on the clustering of the micro-seismicity in the Western High and on a swarm event. As a perspective to future work, an attempt will be made to improve earthquake locations using the dataset from the permanent, cabled, Ocean Bottom Broad-Band Seismometers network operated by KOERI.

Rupin Matthieu

A study of the particular effect of a grid of sub-wavelength resonators on lamb waves

Zaccarelli Lucia

Resolution Limit of Velocity Variations Measured Through Passive Image Interferometry / Spatio-temporal mapping of the crustal damage of a large earthquake through passive image interferometry

Lynda Chehami

Passive acoustic sensing network and imaging

Bayrakci Gaye

Characterization of gas related processes in shallow sediments, based on combined interpretation of acoustic and seismologic datasets

We here present geophysical evidence of gas emissions observed on a multi-parameter dataset from the submerged section of the North Anatolian Fault, within the Sea of Marmara. An acoustic bubble detector (BOB) was deployed on top of the Central High, giving insight on the spatial and temporal variations of the gas emissions. The instrument insonified horizontally the water column over 24 angular sectors of 7° during seven days, providing multiple passes for a given sector. The acoustic data indicate that the gas emission is not a steady process with observed temporal variations of one hour up to 24 hours. A 4-component ocean bottom

Lesage Philippe

Velocity variations in Volcan de Colima, Mexico, associated with a large tectonic earthquake

We present measurements of temporal variations of seismic velocity at Volcan de Colima using seismic noise correlation techniques. About 13 years of continuous recording, from stations located on and outside the volcano, were processed. The daily cross-correlation functions obtained between pairs of stations present high degree of stability over the period of study. This allowed the use of the stretching method to estimate relative velocity variations in the structure. After correcting for the effect of variations of the sampling frequency of the recording system, small velocity fluctuations are obtained that are poorly correlated with the effusive and explosive activity of the volcano. However, a large velocity drop of 1 to 2 % occurred on January 22, 2003 during the MW = 7.4 Tecom√°n earthquake the epicentre of which is located offshore at about 100 km SW from the volcano. The velocity progressively recovered its initial value in about 6 years. Using an approach based on the diffusive behaviour of the coda waves, the region where this velocity variation occurred is located. It is clearly localized in the deep volcanic structure.

Hillers Gregor

Noise-based observation of wave speed variations associated with tidal loading : In-situ acoustoelastic testing

We report on observations of systematic seismic velocity variations in response to tidal deformation. Linear---and potentially nonlinear elastic and/or anelastic---response types and associated material properties from a wide range of Earth materials can be inferred from laboratory acoustoelastic measurements. In this method a continuous low frequency load is applied to a sample and the resulting velocity and hence modulus changes are monitored with simultaneously applied high-frequency pulses. We use a conceptually similar technique to measure the linear response, i.e. the strain sensitivity, of a range of in-situ crustal and fault zone materials. Earth tides constitute the low frequency deformation, relative seismic velocity changes (dv/v) associated with different parts of the tidal strain protocol are resolved using standard ambient noise-based monitoring techniques. This “noise-based in-situ acoustoelastic testing” method is applied in an analysis of ambient seismic wave fields recorded by seismic arrays situated in various tectonic regions covering a range of spatial and temporal scales. We present results from multiple tests to estimate the resolution and temporal robustness of the dv/v signal associated with the deformation protocol, and we discuss the frequency dependence of the signal as a function of ambient wave field properties. We consider the potential to image spatially variable material properties associated with the damage state based on observations of different inter-station dv/v amplitudes in regions with strong lateral heterogeneities (e.g. across fault zones). The new method constitutes a complementary technique to active source experiments for imaging and monitoring rheologic properties and strain sensitivities, and we discuss implications for routine observation and monitoring of response fluctuations and their potential contribution to the detection of transients.

Iglesias Arturo

Attenuation Parameters Obtained by Analysis of Seismic Noise Cross-Correlations

A special case of cross-correlation has been previously proposed to recover information about amplitude behavior of Green‚Äôs functions obtained from the noise of a pair of stations. However, the proposed method implies to avoid temporal normalization as preprocess to maintain information about relative amplitude. An alternative method consists on recover the amplitude information from the ratio of coherent part of the classic cross-correlation function normalized by the very late part of the same function. Following this last strategy, we use data from two dense linear experiments (called MASE & VEOX) consisting on 100 and 50 broadband stations, respectively. These experiments were deployed in Central and South of Mexico between 2005-2007 and 2007-2009. Continuous streams of 100sps were recorded. Hourly segments from the vertical component data of each station were decimated, normalized and whitened before the computation of cross-correlation between pair of stations. These cross correlations are stacked for a large number of hours. The obtained coherent signal was finally normalized with the very late part of the signal. Results show that we were able to recover relative amplitudes between pair of stations and their behaviour can be compared with attenuation laws obtained previously for the region. Note : We presented, partially, this work in a previous meeting (ESC-2012)

Seats Kevin

Seismic structure beneath Yellowstone from ambient noise

We evaluate Rayleigh wave group velocity maps, between 5 and 40 s, inverted for the Yellowstone region using dispersion measurements of vertical component noise correlation functions. We use approximately 180 stations in the area, spanning between 1999 and 2011, including the Transportable Array (TA), as well as the Noise Observatory for Imaging the Subsurface beneath Yellowstone (NOISY) stations. These maps are then used to invert for material properties using a joint inversion scheme with receiver function analysis at those same stations. Additionally, we have begun investigating for time-varying seismic velocities using ambient seismic noise. In active regions, the Noise Correlation Function can change or migrate, reflecting an associated change in velocity structure along that path. We have found percent velocity variations for station paths that sample directly across the Yellowstone magma chamber, with these changes correlating temporally (and spatially to an extent) with other geophysically observed phenomena, such as earthquake swarms and inferred GPS uplift. We have begun attempting to spatially locate these changes, using a time lapsed tomographic technique.

Legrand Denis

Velocity model under Lastarria volcano, Chile using seismic ambient noise

A simple 1D velocity model has been performed on Lastarria volcano, Chile using seismic ambient noise of 4 months of continuous data recorded at 4 broad-band seismometers. A special attention has been done before the summation of 1-hour cross-correlation windows in order to increase the signal/noise ratio. This velocity model has been used to locate volcanic seismicity below the volcano.

Spica Zack

A 2 scales velocity model obtained from seismic noise at Lastarria- Lazufre volcanic zone, Chile-Argentine

A surface wave tomography using two distinct seismic networks below the Lazufre volcanic zone in the central Andes (Chile-Argentine) has been performed using dispersive curves of surface waves reconstructed from cross-correlation of seismic ambient noise. Although no historical eruption have been reported around the Lazufre volcanic zone, some recent InSAR studies highlighted that the region is subject to one of the most important uplift observed on Earth, of about 3 cm/yr. The surface of the deformation yields about 2700 km2 and the source has previously been located between 7 and 17 km beneath the local relief, suggesting that the deformation is a result of a new magma body intrusion. In order to better constrain the source of the deformation and to obtain a velocity image below Lazufre complex, a temporary seismic network, consisted of 18 broad-band stations surrounding the anomaly, was installed during 3 months in 2008, the greatest interstation distance being of 50 km. The results reveal a low velocity zone at 4 to 6 km depth from the surface, located from the center of the maximum deformation to its South East part. A second network distant of about 150 km from the zone has been use and combine to the temporary network allowing us to highlight deeper the volcanic anomaly. Comparing the Lazufre volcanic zone to others uplifting volcanic system as the Uturuncu (Bolivia) can give us some new keys in order to understand the dynamic of the alitplano-puna plateau.

Ruigrok Elmer

Receiver-pair seismic interferometry

With seismic interferometry events can be retrieved between station positions. In the classical form, the events are retrieved by an integration over sources. For a specific dataset, however, the actual source distribution might not be sufficient to approximate the source integral. Yet, there might be a dense distribution of receivers allowing integration over the receiver domain. We rewrite the source integral to an integration over receiver pairs. With this formulation, an event can be retrieved even in the limiting case of only a single source. However, with respect to the classical formulation, an additional stationary-phase analysis is required. Using numerical data, we illustrate how receiver-pair seismic interferometry can be of use in body-wave and surface-wave settings.

Soldati Gaia

Spatio-temporal mapping of the crustal damage of a large earthquake through passive image interferometry

The relative velocity variations possibly associated to large earthquakes can be readily monitored via cross-correlation of seismic noise. Zaccarelli et al. (2011) analyzed more than two years of continuous seismic records from three stations surrounding the epicenter of the April 6, 2009, Mw 6.1 L’Aquila earthquake, observing a clear decrease of seismic velocities likely corresponding to the coseismic shaking. Here we extend their analysis in space, including seismic stations within a radius of 60 km from the mainshock epicenter, and in time, collecting Ô¨Åve years of data for the three nearest stations to the epicenter. Our aim is to investigate how far the crustal damage is visible through this technique, and to detect a potential postseismic recovery of velocity variations.

Faenza Licia

Resolution Limit of Velocity Variations Measured Through Passive Image Interferometry

Seismic noise cross-correlation techniques (or passive image interferometry) are well known to be an appropriate tool for monitoring the relative velocity variations associated to the occurrence of large earthquakes. The measurement resolution depends, from a technical point of view, on the stability of the cross-correlation function, and from a physical point of view on the nature of the geophysical process under study. The possible seismic sources in fact may generate crustal changes with different extension in space and time. In this work we aim to explore the resolution capability of passive image interferometry technique to evaluate the relative velocity variation measurements during seismic swarms, i.e. seismic sequences of small energy. We then took into account the area of the Pollino (in the southern Apennines, Italy), which experienced in the last two years an increase of seismic activity occurring in seismic swarms.

Pellegrini Yves Patrick

Collective-variable description of dynamical dislocation motion

A collective-variable approach is employed to derive a complex-valued equation of motion in that governs the time-evolution of both position and core width of a flat dislocation moving on its slip plane, under the influence of a time-evolving stress. The equation accounts for retarded elastodynamic wave-propagation effects. It applies (with suitable particularizations) to subsonic screw and edge dislocations in an isotropic elastic medium.

Rakotonarivo Sandrine

Structural impedance matrix estimation from a diffuse field

The structural Green’s function can be used to predict the scattered field from an object under any incident field conditions. In this presentation, we derive a method to obtain the structural Green’s function (or the equivalent stiffness or surface impedance matrix) by placing the object in an encompassing noise field and from pressure and normal velocity measurements taken on its surface.

Bonnel Julien

Single-receiver seabed characterization using low-frequency whale vocalization

Natarajan Thulasiraman

Spectral ratio analysis of ambient noise for local site characterization : A case study in Kachchh paleo rift valley

Site response of different geological regimes of Kachchh rift valley, Gujarat, India has been studied using ambient noise recordings. Kachchh is a failed Mesozoic rift where the stress regime had been changed from extensional to compressional after collision of Indian plate with Eurasian plate about 55 Ma. The E-W trending rift hosts many known parallel E-W faults which have generated two large earthquakes (1819, M 7.5 and 2001, Mw 7.6) within a span of 200 years, making it a zone of high seismic risk. We recorded ambient noise data at 55 sites spread across various geological domains in Kachchh, including the sites of extensive liquefaction during the 2001 earthquake, to understand their site responses. We generated H/V spectral ratio curves of ambient noise for each site for a frequency range of 0.1 ‚Äì 15 Hz using the Geopsy tool. We first estimated the predominant frequency (f0) from the prominent peak and then tested their reliability as recommended by SESAME project guidelines. We found that the spatial distribution of f0 reasonably correlated with the regional geology of Kachchh. Further, we observed that many sites did not show a single prominent peak, especially the soft sediment sites exhibited multiple peaks. Further, we carried out hierarchical cluster analysis on H/V curves and grouped similar curves, using which, we intend to develop standard response curves of the different parts of the Kachchh.

Saade Maria

Spacio-temporal variations anisotropy in seismogenic

Seismic anisotropy plays a key role in the study of stress and strain fields in the earth. Potential temporal change of seismic anisotropy can be interpreted as variation of the orientation of cracks in seismogenic zones and thus variation of the stress field. Such temporal changes have been observed in seismogenic zones before and after earthquakes but are still not well understood. In this study, from a numerical point of view, we investigate the variations of the polarization of surface waves in anisotropic media. These variations are related to the elastic properties of the medium, in particular to anisotropy. The technique used is based on the calculation of the whole cross-correlation tensor (CCT) of ambient seismic noise. If the sources are randomly distributed in homogeneous medium, it allows us to reconstruct the Green’s tensor between two stations continuously and to monitor the region through the use of its fluctuations. Therefore, the temporal change of the Green’s cross-correlation tensor enables the monitoring of stress and strain fields. This technique is applied to synthetic seismograms computed in a transversally isotropic medium with horizontal symmetry axis (here after referred to an HTI medium) using a code based on the spectral element method. We designed an experiment in order to investigate the influence of anisotropy on the CCT. In homogeneous, isotropic medium the off-diagonal terms of the Green’s tensor are null. The CCT is computed between each pair of stations and then rotated in order to approximate the Green’s tensor by minimizing the off-diagonal components. This procedure permits the calculation of the polarization angle of quasi-Rayleigh and quasi-Love waves, and to observe the azimuthal variation of their polarization. The results show that even a small variation of the azimuth of seismic anisotropy with respect to a certain pair of stations can induce, in some cases, a large variation in the horizontal polarization of surface waves along the direction of this pair of stations. It depends on the relative azimuth angle between the pair of stations and the direction of anisotropy, on the amplitude of anisotropy and the frequency band of the signal. Therefore, it is now possible to explain the large, rapid and very localized variations of surface waves horizontal polarization observed by Durand et al. (2011) during the Parkfield earthquake of 2004.