Displays

Hamiltonian Monte Carlo (HMC) is a Markov chain Monte Carlo method that exploits derivative information in order to enable long-distance moves to independent models, even when the model space dimension is high (Duane et al., 1987). This feature motivates recent research aiming to adapt HMC for the solution of geophysical inverse problems (e.g. Sen & Biswas, 2017; Fichtner et al., 2018; Gebraad et al., 2020).

Here we present applications of HMC to inverse problems at variable levels of complexity. At the lowest level, we study linear inverse problems, including, for instance, linear traveltime tomography. Though this is not the class of problems for which Monte Carlo methods have been developed, it allows us to understand the important role of HMC tuning parameters. We then demonstrate that HMC can be used to obtain probabilistic solutions for two important classes of inverse problems: 2D nonlinear traveltime tomography and 2D elastic full-waveform inversion. In both scenarios, no super-computing resources are needed for model space dimensions from several thousand to ten thousand.

By far the most important, but also most complex, tuning parameter in HMC is the mass matrix, the choice of which critically controls convergence. Since manual tuning of the mass matrix is impossible for high-dimensional problems, we develop a new HMC flavour that tunes itself during sampling. This rests on the combination of HMC with a variant of the L-BFGS method, well-known from nonlinear optimisation. L-BFGS employs a few Monte Carlo samples to compute a matrix factorisation LL^Twhich dynamically approximates the local Hessian H, while the sampler traverses model space in a quasi-random fashion. The local curvature approximation is then used as mass matrix. Following an outline of the method, we present examples where the auto-tuning HMC produces almost perfectly uncorrelated samples for model space dimensions exceeding 10⁵.

References

[1] Duane et al., 1987. "Hybrid Monte Carlo", Phys. Lett. B., 195, 216-222.

[2] Sen & Biswas, 2017. "Transdimensional seismic inversion using the reversible-jump Hamiltonian Monte Carlo algorithm", Geophysics, 82, R119-R134.

[3] Fichtner et al., 2018. "Hamiltonian Monte Carlo solution of tomographic inverse problems", Geophys. J. Int., 216, 1344-1363.

[4] Gebraad et al., 2020. "Bayesian elastic full-waveform inversion using Hamiltonian Monte Carlo", J. Geophys. Res., under review.

How to cite: Fichtner, A., Gebraad, L., Boehm, C., and Zunino, A.: Auto-tuning Hamiltonian Monte Carlo, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7735, https://doi.org/10.5194/egusphere-egu2020-7735, 2020.

BEL1D has been newly introduced to the community as a viable algorithm for the stochastic interpretation of geophysical data in the form of 1D geological models. It relies on a simplified version of the Bayesian problem in reduced space called Bayesian Evidential Learning. However, the method is closer to machine learning than classical McMC approaches since it can be separated into a learning process followed by a prediction part. The learning phase consists in constituting statistical relationships between models parameters and geophysical data from a training set of numerical models. The prediction phase then samples the previous relationships according to field data. Compared to other stochastic methods such as McMC, BEL1D as key advantages: 1) it converges easily as long as the prior is consistent with the unique input parameter being the size of the training set, 2) every model in the posterior is drawn independently, making it easy to trace back their origin, 3) the CPU times are similar to McMC, but the method can be fully parallelized and the learning process can be done before data acquisition, leading to quasi instantaneous prediction of the posterior. BEL1D already has led to successful applications on surface nuclear magnetic resonance data as well as dispersion curves from surface waves analysis. Nonetheless, the method is not limited to those two examples and can be implemented for any 1D geophysical method as long as a forward model is provided. Currently, the method is implemented for blocky imaging but will be extended to non-blocky models in the future. The open-source codes are readily available.

How to cite: Michel, H., Nguyen, F., and Hermans, T.: BEL1D: 1D imaging using geophysical data in the framework of Bayesian Evidential Learning, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-9190, https://doi.org/10.5194/egusphere-egu2020-9190, 2020.

The full-waveform inversion (FWI) of surface waves, including both Rayleigh and Love waves, is becoming increasingly popular for near-surface characterizations. Due to the high nonlinearity of the objective function and a huge amount of data, FWI may converge towards a local minimum and is usually computationally expensive. To overcome these problems, we reformulate FWI under a multi-objective framework and propose a random objective waveform inversion (ROWI) method for surface-wave characterization. We use three objective functions: the classical least-squares (l₂) waveform difference, the envelope difference, and the difference in the FK spectra. At each iteration, we randomly choose one shot and randomly assign one of the three objective functions to this shot. We only update the model with one iteration using a preconditioned steepest descent algorithm to optimize the currently assigned objective function. Therefore, ROWI has high freedom in exploring the model and objective spaces.
We use a synthetic example to compare the performance of ROWI with conventional FWI approaches. ROWI converges to better result compared to the conventional FWI approaches, while some of the conventional FWI approaches are trapped at local minima and fail to reconstruct reasonable results. We also apply ROWI to a field data acquired in Rheinstetten, Germany. The main geological feature, a refilled trench, is successfully reconstructed in the ROWI result. The reliability of the ROWI result is also proven by a migrated GPR profile. Overall, both synthetic and field-data examples show that ROWI is computationally more efficient, less dependent on the initial model, and more robust compared to conventional FWI approaches.

How to cite: Pan, Y., Gao, L., and Bohlen, T.: Random objective surface-wave waveform inversion, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-3910, https://doi.org/10.5194/egusphere-egu2020-3910, 2020.

Surface Wave Tomography (SWT) is a well-established technique in global seismology: signals from strong earthquakes or seismic ambient noise are used to retrieve 3D shear-wave velocity models, both at regional and global scale. This study aims at applying the same methodology to controlled source data, with specific focus on 3D acquisition geometries for seismic exploration. For a specific frequency, travel times between all source-receiver couples are derived from phase differences. However, higher modes and heterogeneous spatial sampling make phase extraction challenging. The processing workflow includes different steps as (1) filtering in f-k domain to isolate the fundamental mode from higher order modes, (2) phase unwrapping in two spatial dimensions, (3) zero-offset phase estimation and (4) travel times computation. Surface wave tomography is then applied to retrieve a 2D phase velocity map. This procedure is repeated for different frequencies. Finally, individual dispersion curves obtained by the superposition of phase velocity maps at different frequencies are depth inverted to retrieve a 3D shear wave velocity model.

How to cite: Barone, I., Kästle, E., Strobbia, C., and Cassiani, G.: Seismic Surface Wave Tomography on dense 3D active data, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7601, https://doi.org/10.5194/egusphere-egu2020-7601, 2020.

The Bangong-Nujiang suture zone, located in the central Tibet, is one of several important geological boundaries in Qinghai-Tibet plateau. Abundant researches have been performed and most of them focused on deep tectonic structure and its dynamic mechanism through recent geophysical projects such as INDEPTH-III, Hi-CLIMB, ANTILOPE, SinoProbe, etc. (Zhao Wenjin et al., 2008; N´abelek et al. 2009; Gao Rui, et al., 2013;Zhao Junmeng et al. 2014; He Rizheng et al., 2014; Xu Qiang et al., 2017; Shang Xuefeng et al., 2017; Davlatkhudzha et al.,2018). Near-surface velocity study can not only obtain the physical parameters such as Vp and Vs in the area, but also improve seismic image quality of deep structure (Zhao Lingzhi et al., 2018). However, the velocity information obtained from passive seismic stations using either receiver function or ambient noise tomography is not enough to elaborate the near surface velocity structure of the Bangong-Nujiang suture zone. Besides, the active-source seismic reflection data usually doesn’t have sufficient offset density at near surface which poses a challenge to conventional near-surface velocity analysis methods.

This study makes full use of surface waves and first breaks to obtain near-surface P- and S-wave velocities based on a 2D deep seismic reflection survey data which was acquired by SinoProbe project in 2009 . We adopt the method of superposition of surface waves in common receiver domain to generate high quality F-K spectrum which enables us to obtain fundamental-order and high-order dispersion curves. First, a 2D layered model with an irregular topography was built and the 2D elastic finite difference modeling was executed to generate 161 synthetic seismic shot gathers which mimicking the actual acquisition geometry. These gathers contain surface waves, refractions, reflections and multiples energy, and the maximum offset is about 18 km. It is shown that the F-K spectrum quality has been improved for each receiver station using superposition of surface waves in the F-K domain by adding more shots. The S-wave velocity inverted from dispersion curves showed good agreement with the synthetic model. Second, high quality F-K spectrum generated from the above method enabled us to pick both fundamental and 1^st order dispersion curves from the SinoProbe field data. The S-wave velocity was generated using three methods: 1) empirical equations based on dispersion curves; 2) fundamental order dispersion curves inversion; and 3) both fundamental and 1^st order dispersion curves inversion. Results show that using higher order dispersion curves can generate a more reliable near-surface model. Third, first breaks were picked up to 18 km offset and diving wave tomography was applied to derive near-surface P-wave velocity from abundant first break information. It is shown that there is an excellent correlation between P- and S-wave velocities, the bottom of basin is clearly revealed, and over-thrusts are identified accordingly which is consistent with field geological survey in the middle segment of Bangong-Nujiang suture zone.

This study was financially supported by the CAGS Research Fund (grant YWF201907), and the National Natural Science Foundation of China (grant 41761134094). Data sources: SinoProbe-02 Project.

How to cite: Zhang, H., He, R., and Liu, Z.: Near-surface velocity structure study using surface waves and first breaks in the middle segment of the Bangong-Nujiang suture zone, Tibetan Plateau, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6253, https://doi.org/10.5194/egusphere-egu2020-6253, 2020.

Unaweep Canyon (Western Colorado, US) is an enigmatic alpine landform and hypothesized to represent a partially exhumed paleo valley which was glacially over-deepened in the late Paleozoic. Processing and interpretation of recently acquired 2D seismic reflection and refraction data support the concept of glacial over-deepening and indicate maximum bedrock depths of about 550 meters. Additionally, pronounced reflectors are observed within the sedimentary infill. The seismic data have also been subjected to surface wave analysis revealing a significant increase of the Vp/Vs ratio below a shallow (50 – 150 m depth) intra-sedimentary reflector. A large Vp/Vs ratio can be caused by both saturation and poor consolidation of dry low-porosity materials (e.g. dry sands).

To investigate the potential occurrence of an aquifer associated with this interface, a high-density/long-offset electrical resistivity survey was conducted in fall 2019 along the seismic line. The maximum offset is 915 m at an electrode spacing of 5 meters, aiming at reaching depths of investigations between 150 and 200 meters. Inversion of the ERT data was initially conducted by means of smoothness-constrained algorithms. The imaging results revealed consistent structures with those resolved through seismic methods, at least within the required depth of investigation between 150 – 200 m. Furthermore, improvements in the resolution of the ERT imaging results was investigated after the inclusion of seismic interfaces as structural constraints in the inversion of the data. The comparison of the two approaches permitted to improve the interpretation of the ERT imaging results, which indicate low resistivities in the zone of high Vp/Vs ratios and thus strengthen the aquifer hypothesis. We present an integrated interpretation based on seismic structure, resistivity distribution, Vp and Vs velocities, and a distant well core. In a larger context, the results provide new insights on the subsurface hydrology in this arid part of the continental US as well as on the significance of multi-valued datasets for the interpretation and characterization of aquifers.   

How to cite: Behm, M., Flores-Orozco, A., Chwatal, W., and Soreghan, G. S.: Hydrologic characterization of an alpine valley infill through integration of ERT, active seismic and active/passive surface wave interferometry (Unaweep Canyon, US), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6199, https://doi.org/10.5194/egusphere-egu2020-6199, 2020.

The 1D layered inversion of surface wave dispersion data is a powerful tool to characterize the vertical distribution of S-wave velocity. Its applications span from seismology to geotechnical engineering, going through exploration geophysics. As many others, also this non-linear inverse problem is considerably ill-posed. Thus, in the Tikhonov’s regularization framework, the associated non-uniqueness and instability of the solution with respect to the data and their uncertainty can be tackled by including prior information in the inversion process. However, for the case of the gradient-based deterministic inversion problem, only constraints enforcing smooth spatial variations of the S-velocities have been used, even when blocky targets were expected. This, clearly, generates results that might fit the observed data, but that are often not compatible with other sources of information. On the other hand, probabilistic approaches can be used to properly map the model space; however, they are still very computationally expensive to be used routinely, or to be easily integrated in a multi-physical inversion procedure involving other geophysical methods.

Our goal is to combine computer efficiency, capability of integration with other geophysical methods, and some exhaustiveness regarding the non-uniqueness of the inverse problem. For this, we developed a coherent set of tools for the deterministic inversion of dispersion curves that is capable of applying a quite large spectrum of constraints. This includes, for example, vertically and laterally constrained inversions with different levels and kinds of regularization (sharpness and/or smoothness). In this study, we evaluate the capabilities and the possible limitations of the different regularization approaches on various datasets.

How to cite: Guillemoteau, J., Vignoli, G., and Barreto, J.: Comparison of different regularization schemes for the 1D laterally constrained inversion of seismic surface wave data, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11061, https://doi.org/10.5194/egusphere-egu2020-11061, 2020.

In the recent years, the diffracted wavefield has gained increasing attention in the field of applied seismics. While classical seismic imaging and inversion schemes mainly focus on high-amplitude reflected measurements, the faint and often masked diffracted wavefield is neglected or even treated as noise. In order to be able to extract depth-velocity models from seismic reflection data, sufficiently large source-receiver offsets are needed. However, the acquisition of such multi-channel seismic data is expensive and often only feasible for the hydrocarbon industry, while academia has to cope with low-fold or zero-offset data. The diffracted wavefield is the key for extracting depth velocities from such data, as the moveout of diffractions – in contrast to reflections – can be measured in the zero-offset domain. Recently, we have demonstrated on multi-channel, single-channel and passive seismic data that by means of wavefront tomography depth-velocity models can be retrieved solely based on diffractions or passive seismic events along with the localizations of these scatterers. The input for wavefront tomography are so-called wavefront attributes, which can be extracted from the data in an unsupervised fashion by means of coherence analysis. In order to obtain the required diffraction-only data, we use a recently proposed scheme that adaptively subtracts the high-amplitude reflected wavefield from the raw data. Due to their most common acquisition geometry, most ground-penetrating-radar (GPR) data inherently lack offsets. In addition, GPR data generally contain a rich diffracted wavefield, which in turn contains information about sought-after structures, as diffractions are caused by small-scale heterogeneities such as faults, tips or edges. In this work, we show an application of the suggested workflow – coherence analysis, diffraction separation and diffraction wavefront tomography – to GPR data acquired at a glacier, resulting in a depth-velocity model and the localizations of the scatterers, both obtained in a fully unsupervised fashion. While the resulting  velocity model may be used for depth migration of the raw data, the localizations of the scatterers may in addition provide important information on the inner structure of the glacier in order to, for instance, localize water intrusions or fractures.

How to cite: Bauer, A., Schwarz, B., Delf, R., and Gajewski, D.: Application of diffraction wavefront tomography to GPR data from a glacier, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17349, https://doi.org/10.5194/egusphere-egu2020-17349, 2020.

Understanding subsurface water flow is important as it e.g. controls contaminant transport, has an impact on the amount of aquifer recharge, and can be used for storm water management purposes. However, there do not exist many methods that can observe the water flow in the field. Furthermore, the flow patterns can be very diverse due to the complex geological conditions, e.g. faults, fractures, and heterogeneous permeability of the subsurface formations. In order to map the subsurface water flow in a chalk formation, we performed a water injection experiment in the Rørdal Quarry, Northeast Denmark. A total water volume of 700 liters was injected via a 50 cm deep hole within 8 hours. Around the injection hole, we conducted time-lapse GPR measurements along 6 inlines and 6 crosslines. Seven measurements campaigns were performed over an eight-hour time period. We analyze the time-lapse GPR reflection sections in order to investigate the variations of the different measurements. Initially, we subtract the repeated measurements and baseline measurements, which shows that some survey lines have clear changes after water injection, while others only show very small or no changes. To verify the differences, we pick travel times of selected horizons in the time-lapse data and compare them (cf. Truss et al., 2007; Allroggen et al., 2015). This analysis highlights the travel time variations imposed by the injected water. Moreover, we perform correlation analysis of the measurements before and after water injection. The correlation coefficients show relatively small values on the lines that exhibit clear differences, further confirming the differences caused by the water infiltration. Initial integrated analysis of the different results shows that the water mainly flows towards the southeast from the injection hole. This is consistent with the orientation of the fracture system observed in the reflection GPR profiles, indicating that the water flow is primarily controlled by the fractures.

[1] Truss, S., Grasmueck, M., Vega, S., and Viggiano, D. A. 2007, Imaging rainfall drainage within the Miami oolitic limestone using high-resolution time-lapse ground-penetrating radar, Water Recourses Research, 43, W03405.

[2] Allroggen, N., Schaik, N.V., and Tronicke, J. 2015, 4D ground-penetrating radar during a plot scale dye tracer experiment, Journal of Applied Geophysics, 118, 139-144.

How to cite: Yuan, H., Zibar, M. C. L., and Nielsen, L.: Subsurface water flow detection by time-lapse reflection GPR data, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4493, https://doi.org/10.5194/egusphere-egu2020-4493, 2020.

Seismic images provided by standard Reverse Time Migration are usually contaminated by artefacts associated with the migration of multiples.

Multiples can corrupt seismic images by producing both false negatives, i.e. by destructively interfering with primaries, and false positives, i.e. by focusing energy at unphysical interfaces. Free-surface multiples particularly affect seismic images resulting from marine data, while internal multiples strongly contaminate both land and marine data. Multiple prediction / primary synthesis methods are usually designed to operate on point source gathers, and can therefore be computationally  demanding when large problems, involving hundreds of gathers, are considered.

In this contribution, a new scheme for fully data-driven retrieval of primary responses of plane-wave sources is presented. The proposed scheme, based on convolutions and cross-correlations of the reflection response with itself,  extends a recently devised Marchenko point-sources primary retrieval method for to plane-wave source data. As a result, the presented algorithm allows fully data-driven synthesis of primary reflections associated with plane-wave source data. Once primary plane-wave responses are estimated, they are used for multiple-free imaging via standard reverse time migration. Numerical tests of increasing complexity demonstrate the potential of the proposed algorithm to produce multiple-free images only involving the migration of few datasets.

The plane-wave source primary synthesis algorithm discussed in this contribution could then be used as an initial and unexpensive processing step, potentially guiding more expensive target imaging techniques. Moreover, the method could be applied to large 3D problems for which standard methods are prohibitively expensive from a computational point of view.

How to cite: Meles, G. A., Zhang, L., Thorbecke, J., Wapenaar, K., and Slob, E.: Data-driven synthesis of primary plane-wave responses, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8606, https://doi.org/10.5194/egusphere-egu2020-8606, 2020.

The imaging of seismic reflection data provides a powerful high-resolution method for studying volcano structure and fluids presence. The shallow structure of the Solfatara crater, a surface marker of deep magmatic activity inside Campi Flegrei caldera (Southern Italy), is characterized in terms of seismic profile and attributes. The main contribution of this work is to provide a detailed and improved seismic reflection image of the Solfatara crater and the identification of gas accumulation. The profiles are deployed along the NNE-SSW directions, the first, and the second orthogonal to the last. The two profiles are 400 m long acquired during the active experiment RICEN (Repeated Induced Earthquake and Noise) performed in the framework of the EU project MEDSUV between May and November 2014. Pre-stack processing of the seismic data has been performed in order to remove the noisy traces, low-frequency noise and reduce the ground roll phases. A very detailed velocity analysis for the NMO correction has been performed with the integration of information derived from the Vp velocity model previously obtained by the non-linear Bayesian technique. After having applied residual statics and DMO corrections, the CMP gathering, the post-stack Kirchhoff migration technique was performed to produce the final seismic profiles in time and depth. Once having obtained the post-stack migrated imaged, the energy, root mean square, envelope and sweetness attributes were computed for defining the maximum and minimum value of amplitude zones. In addition, other attributes as the time-gain attribute in order to interpret the deep reflectors and the variance attribute to define the faults, discontinuities, and chaotic zones have been evaluated. To enhance fluids identification the Amplitude Versus Offset (AVO) variation technique has been further applied to identify the gas zone in the explored sections. By integrating all information from the original seismic profile, seismic attributes and geophysical investigation relative to the Solfatara volcano, the multi-2D image presents the fluids trapped in the Solfatara crater at depths between 10 to 50 m below the surface of the crater and their migration pathways up to 150 meters depth.

How to cite: Gammaldi, S., Ismail, A., Chiuso, T., and Zollo, A.: The multi-2D seismic imaging of the Solfatara Volcano, Italy, inferred by seismic attributes., EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-16478, https://doi.org/10.5194/egusphere-egu2020-16478, 2020.

The karst lakes of the sparsely-populated Lacandon Forest in Chiapas, southern Mexico, and their associated sediment infill are attracting increasing attention as high-resolution and continuous environmental and climate archives. To evaluate the information stored in the sediments, paleolimnologists retrieve sediment cores and analyze multiple biological and non-biological indicators. Our geophysical measurements presented here were motivated by the need to determine coring locations providing continuous sediments records from a total of four lakes of the Lacandon Forest. Therefore, we mapped the sediment thickness on the lake floor by applying seismic, electrical, and electromagnetic methods. The measurements were carried out with floating devices – and, after the sudden drainage of two of the studied lakes, complemented by measurements on the exposed lake floor.

During a first campaign in March 2018 when lakes were filled, we collected seismic data with a sub-bottom profiler (SBP). Furthermore, we collected transient electromagnetic (TEM) data with a floating measuring device to investigate the potential of the method for the determination of sediment thicknesses as an alternative to seismic methods. After the lake-level maximum that coincided with the first campaign, the water levels of two of the studied lakes dropped dramatically by July 2019, leaving lake Metzabok (maximum depth ~15 m) dry and lake Tzibaná (~70 m) with a water level decreased by approx. 30 m. In October 2019, when lake levels were still low, we conducted a second survey covering the dry lake floor of lake Metzabok and some dry parts of lake Tzibaná. During this second campaign, we collected electrical resistivity tomography (ERT), induced polarization (IP), and seismic refraction tomography (SRT) data along selected lines of the 2018 survey.

Our 2018 results from the water-borne survey show that sediment thickness estimates from seismic (SBP) and electrical (TEM) data agree well for water depths up to 20 m and sediment thicknesses ranging from 2 m to 10 m. The 2019 data collected on the dry lake floor confirms the findings of the first campaign and – due to the smaller distance between measuring devices and target – results in a more detailed picture of sediments and the underlying limestone bedrock.

How to cite: Bücker, M., Pérez, L., Flores Orozco, A., Gallistl, J., Steiner, M., Aigner, L., Hoppenbrock, J., Morales Barrera, W., Pita de la Paz, C., García García, E., Razo Pérez, J. A., Buckel, J., Hördt, A., and Schwalb, A.: Water- and land-borne geophysical measurements before and after the sudden drainage of large karst lakes in southern Mexico, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11031, https://doi.org/10.5194/egusphere-egu2020-11031, 2020.

This paper collects 43,225 absolute first arrival P wave arrival times and 422,956 high quality relative P arrival times of 6,390 events occurred in Yangjiang and its adjacent area from Jan, 1990 to Aug, 2019, these seismic data is recorded by 49 stations from Guangdong seismic network, Guangxi seismic network and Hainan seismic network. Based on the seismic data above, we simultaneously determine the crustal 3D P wave velocity structure and the hypocenter parameters of 6255 events in Yangjiang and its adjacent area by applying Double-Difference seismic tomography. The result shows that, shallow P wave velocity in Yangjiang area is higher due to the thinner sedimentary layer and widely exposed Yanshanian granite, Indosinian granite and Cambrian metamorphic rocks. There are obvious correspondences between the distribution of shallow velocity and fault structure as well as geological structure. A wide range of low velocity anomaly exists in 20km depth, which verifies the low velocity layer in the middle crust at Yangjiang area of South China continent. The velocity image from land to ocean in 30km depth shows low velocity in NW side and high velocity in SE side, which verifies the characteristic of crust thinning in South China coastal continent. The NEE seismic belt from Yangbianhai to Pinggang is speculated to locate in a buried fault of southwest segment of Pinggang fault. The buried thrust fault is a N78°E strike fault, dip to NW with a dip angle of 85 °. In addition, the buried fault locates in the abnormal junction of high velocity on the NW side and low velocity on the SE side, which reflects the tectonic activity characteristic of NW plate uplifting and SE plate declining from Miocene period. The characteristic of activity in the buried fault shows thrust movement with a small strike slip component, which is consistent with the focal mechanism of M4.9 earthquake occurred in 2004.

How to cite: Wang, X., Deng, Z., Ye, X., and Wang, L.: The Study of Crustal Velocity Structure Characteristics in Yangjiang Area of South China, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4319, https://doi.org/10.5194/egusphere-egu2020-4319, 2020.

Confirming the permanent containment is a key challenge for the storage of CO₂ in deep underground reservoirs. Faults in the cap rock of such reservoirs are potential flow paths for the CO₂ to escape. Our decametre-scale experiment at the Mont Terri Rock Laboratory aims to better understand mechanisms of CO₂ leakage trough a fault, and to test strategies to monitor the propagation of CO₂-saturated water through faults.

Two boreholes were drilled through the main fault in Mont Terri with packer-intervals dedicated to fluid-injection and hydraulic/geochemical monitoring. Another five boreholes in the close surrounding were equipped with various instruments for geotechnical and geophysical observations. During the first phase of the experiment, the hydraulic response of the fault was characterized with injections of formation water in a step-up mode at pressures up to 6.0 MPa. The second phase, which was still on-going at the time of the abstract submission, consists of a long-term (several months) injection of CO₂-saturated formation water at a constant head of 4.5 MPa, which is below the fault opening pressure. All injection activities were monitored with active seismic measurements, along with a comprehensive set of hydraulic-, mechanical-, geochemical- and other geophysical surveys. We will present the active seismic imaging results from the step-up injection test and compare them with the other surveys. Additionally, preliminary results will be shown acquired during the long-term injection of CO₂-saturated formation water into the fault.

How to cite: Grab, M., Zappone, A., Rinaldi, A. P., Hellmann, S., Wenning, Q., Roques, C., Obermann, A., Madonna, C., Guglielmi, Y., Nussbaum, C., Maurer, H., and Wiemer, S.: Active seismic monitoring of CO2-saturated brine injection into a fault (CS-D experiment in the Mont Terri Rock Laboratory), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-21588, https://doi.org/10.5194/egusphere-egu2020-21588, 2020.

Given the sparsity of geophysical data it is useful to rely on prior information on the expected geological patterns to constrain the inverse problem and obtain a realistic image of the subsurface. By using several examples of such patterns (e.g. those obtained from a training image), deep generative models learn a low-dimensional latent space that can be seen as a reparameterization of the original high-dimensional parameters and then inversion can be done in this latent space. Examples of such generative models are the variational autoencoder (VAE) and the generative adversarial network (GAN). Both usually include deep neural networks within their architecture and have shown good performance in reproducing high-dimensional structured subsurface models. However, they both use a highly nonlinear function to map from latent space to the original high-dimensional parameter space which hinders the optimization of the objective function during inversion. Particularly, such nonlinearity may give rise to local minima where gradient-based inversion gets trapped and therefore fails to reach the global minimum. GAN has been previously used with gradient-based inversion in a linear traveltime tomography synthetic test where it was shown to often fail in reaching a consistent RMSE (compared to the added noise) because optimization converges to local minima. On the other hand, inversion with MCMC and GAN was shown to reach acceptable RMSE values. When applicable, however, a gradient-based inversion is preferred because of its lower computational demand. We propose using VAE together with gradient-based inversion and show that optimization reaches lower RMSE values on average compared to GAN in a linear traveltime tomography synthetic case. We also compare the subsurface models that are generated during the iterations of the optimization to explore the effect of the different latent spaces used by GAN and VAE. We identify a trade-off between a strict following of the patterns and getting trapped in local minima during optimization, i.e. VAE seems to be able to break some continuous channels in order to not get trapped in local minima whereas GAN does not break channels. Finally, we perform some synthetic tests with nonlinear traveltime tomography and show that gradient-based inversion with VAE is able to recover a similar global structure to the true model but its final RMSE values are still far from the added noise level.

How to cite: Lopez-Alvis, J., Laloy, E., Hermans, T., and Nguyen, F.: Constraining gradient-based inversion with a variational autoencoder to reproduce geological patterns, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-19576, https://doi.org/10.5194/egusphere-egu2020-19576, 2020.

Despite strong evidences that are visible at the surface that suggests the presence of buried structures, sometimes, both the GPR and magnetic data do not allow to clearly about the presence of these structures. Usually, this lack of perceptibility is due to the physical and chemical conditions of the medium that produces an increasing of background noise and masks the useful information. This causes a decrease in the signal-to-noise ratio of the data, preventing a good assessment about the existence of buried structures at subsurface.

Nevertheless, we believe that the recorded signal of both methods has the useful part of the signal hidden. Data fusion techniques are widely used in brain tumour detection in medicine by combining data from different clinical exams, both with low perceptibility.

This work presents an approach that allows using advanced fusion algorithms to combine geophysical data from GPR-3D and magnetics. This creates an enhanced image from both datasets with better quality than the individual images from each method.

The data fusion approach is performed through the combined use of 2D Discrete Wavelet Transform, Multiresolution Singular Value Decomposition and Image Gradient. This scheme allows us to select the useful information to obtain a higher quality and sharper fused image using the best of input datasets. The geophysical data fusion was successfully tested on three datasets, with different levels of perceptibility: high, intermediate and low.

Acknowledgment: This work is co-funded by the ICT Project (UID/GEO/04683/2019) with the reference POCI-01-0145-FEDER-007690, by the Project SFRH/BSAB/143063/2018 (FCT) and by the INTERREG 2014-2020 Program, through the "Innovación abierta e inteligente en la EUROACE" Project, with the reference 0049_INNOACE_4_E.

How to cite: Oliveira, R. J., Caldeira, B., Teixidó, T., and Borges, J. F.: Advanced fusion of geophysical data through combined use of 2D Discrete Wavelet Transform and Multiresolution Singular Value Decomposition applied to GPR-3D and magnetic data, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-20500, https://doi.org/10.5194/egusphere-egu2020-20500, 2020.

In recent decades, geoelectrical methods have played a very important role in near-surface investigation. The most widely used of these methods is electrical resistivity tomography (ERT). Regardless of the forward and inversion algorithms used, the original data collected from a survey is the most important factor for quality of the resulted model. However, 3D electrical resistivity survey design continues to be based on data sets recorded using one or more of the standard electrode arrays. There is a recognized need for the 3D survey design to get better resolution using fewer data. Choosing suitable data from the comprehensive data set is a great approach. By reasonable selecting, better resolution can be obtained with fewer electrodes and measurements than conventional arrays. Previous research has demonstrated that the optimized survey design using the 'Compare R' method can give a nice performance.

This paper adds target-oriented selection and modified the original 'Compare R' method. The survey design should be focused on specific target areas, which need a priori information about the subsurface properties. We select electrodes and configurations as the target set by the comprehensive set firstly which meets the requirements of the target area. The number of measurements and electrodes is much less than the comprehensive set and the model resolution matrix takes less time to calculate. At the next step for rank, we calculate the sensitivity matrix of the target set only once and then calculate the contribution degree of each measurement separately from it. The time of iterative calculation of the resolution matrix when measurements set changing is less than the original method.

The traditional method of evaluating RMS is not appropriate for comparing the quality of collected data by different survey designs. SSIM (structural similarity index) gives more reliable measures of image similarity better than the RMS. The curves of SSIM values in three dimensions and the average SSIM are given as quantitative comparisons. Besides, the frequency of electrodes utilized given to guides on selecting the highest used electrodes. Finally, the curves of the average relative resolution S and the number of electrodes as the number of measurements increase are given, which proves the method works effectively.

The results show the significance of using target-oriented optimized survey design, as it selects fewer electrodes and arrays than the original CR method. Also, it produces better resolution than conventional arrays and takes less calculation time. 3D SSIM, frequency of electrodes used, the relationship between average relative resolution, number of electrodes and number of measurements, these quantitative comparison methods can effectively evaluate the data collected in various survey designs.

How to cite: Jiang, L., Tian, G., Wang, B., and Abd El-Raouf, A.: Target-oriented optimized survey design and quantitative comparison for 3D electrical resistivity tomography, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-3233, https://doi.org/10.5194/egusphere-egu2020-3233, 2020.

Surface seismic methods are among the most popular, widely accepted, geophysical methods for near-surface characterization. The most practical and effective way to perform in-situ measurements and data processing using different seismic methods as are seismic refraction, seismic reflection and MASW method in an integrated approach is presented in this paper. Each method has some advantages and limitations, but their application in an integrated approach provides higher accuracy in subsurface modeling. The same seismic equipment and, in most of the cases, the same acquisition parameters were used, enabling time and cost effective survey for subsurface characterization. The choice of these parameters was not random. Experimental research by use of the above-mentioned seismic methods was carried out in a long period in order to define the optimal parameters for successful application of an integrated technique in future research. During this survey, particular attention was paid to the influence of the acquisition parameters on the dispersion image resolution in the MASW surveys and extraction of an effective dispersion curve.

The results of the performed surveys at characteristic locations in R. North Macedonia are presented to show the efficiency of the combined methods approach.

How to cite: Gjorgjeska, I., Sheshov, V., Edip, K., and Dojchinovski, D.: Efficiency Of Integrated Seismic Methods Approach To Near-Surface Characterization, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17489, https://doi.org/10.5194/egusphere-egu2020-17489, 2020.

The present work consists in imaging salt bodies from earth subsoil in the context of Reverse Time Migration (RTM) algorithm. The study of salt domes is economically important because they form a natural trap for hydrocarbons. For instance, more than a half of the hydrocarbon reserves that still exist today are related to salt bodies.

However, seismic images coming from strong salt tectonics area, are contaminated with spurious signal, like multiple events. Therefore, it is important to know how to treat and filter multiples in order to have seismic images that are geologically interpretable.

For this purpose, we solved the forward 3D elastic seismic wave equations using high order finite differences. The earth parameters come from 3D velocity and density models in a salt tectonic region in the North Gulf of Mexico. To obtain the imaging condition we compute the sensitivity kernels by using the adjoint solution of wave equation and by applying checkpointing. We tested this algorithm with simultaneous and separated sources. Fluid - solid interfaces at the ocean bottom is introduced, interfaces are well retrieved at large offsets.

Furthermore, we applied CPML absorbing boundaries, and replace also free surface conditions for absorbing boundaries to attenuate free surface multiples. The images we obtained from sensitivity kernels are easily interpretable. The calculations were performed on CALMIP supercomputing platforms in Toulouse France. 

How to cite: Abreu, J., Martin, R., and Darrozes, J.: 3D high order seismic imaging in Gulf of Mexico in the context of RTM algorithm using adjoint-based methods., EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-90, https://doi.org/10.5194/egusphere-egu2020-90, 2020.

We solve elastic waves equations in 2D/3D using a fourth order in time and space finite volume method based on exact Riemann solver and wave limiters and designed particularly to capture shock waves.  We validate our code by comparing our results with spectral elements solutions (SPECFEM). 

The goal is to detect the levels of fluid saturations in complex rheological media at different scales (near surface or crustal scale) by homogenizing the rheological laws for different frequency contents of the signal sources. For this purpose, at high frequencies, attenuation and/or non linear effects must be added to the stress-strain relations that affect the apparent/effective seismic wave velocities due to unconsolidated, granular or damaged and fractured nature of solid media.

In our present studies, we show first how we include the viscoelastic models using auxiliary differential equation approach as in Martin et al. 2019. Then, we show at two different scales (laboratory and natural near surface) how we are able to reproduce real seismic data for granular and porous media in the range of 300-4000 Hz for a 1m configuration experimental setup and how the attenuation and homogeneized seismic velocities of the porous matrix correlate to gravity laws and can explain the recorded signals. Recorded signals are compared to the numerical solutions for three different vertical pore pressure gradients due to the injection of gas.  

After validating the code on this controlled experiment we extend our methods at the scale of a 100m natural site in the Orgeval basin in the Parisian region/ France where we want to model the variations of the seismic velocities in the first tens of meters close to the near surface due to the presence of water contents. The idea is to be able to monitor the fluid contents of the ground in the neighborhood of the rivers and the fluid exchanges between nappes, rivers and the underground at the different periods of time (weekly, monthly and annualy). The Vp/Vs ratios reaching high values around 2 due to the water presence, and the Vp and Vs models being obtained by ray-tracing based first arrival time inversions.

The data are provided by the PIREN-Seine program in the last three years.

Complex non-linearity responses of the medium  can be modelled with our codes and also be extended to  damaged faults that can be activated and trigger earthquakes due to bulk and shear moduli decrease.

How to cite: Asfour, K., Martin, R., and El Baz, D.: Numerical modeling of attenuation and non linear effects in computational seismology, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-91, https://doi.org/10.5194/egusphere-egu2020-91, 2020.

Large quantities of fluids are predicted to be expelled from compacting sediments on subduction margins. Fluid expulsion is thought to be focussed, but its exact locations are usually constrained on very small scales and rarely can be resolved using velocity images obtained from traditional velocity analysis and ray-based tomography because of their resolution and accuracy limitation. However, with recent advancement in computing power, the full waveform inversion (FWI) is a powerful alternative to those traditional approaches as it uses phase and amplitude information contained in seismic data to yield a high-resolution velocity model of the subsurface.

Here, we applied elastic FWI along an 85 Km long 2D multichannel seismic profile on the southern Hikurangi margin, New Zealand. Our processing sequence includes: (1) downward continuation, (2) 2D traveltime tomography, and (3) full waveform inversion of wide-angle seismic data. We will present the final high-resolution velocity model and our interpretation of the fluid flow regimes associated with both the deforming overriding plate and the subducting plate.

How to cite: Djeffal, A., Pecher, I., Singh, S., and Kaipio, J.: High Resolution Imaging of the South Hikurangi Subduction Zone, New Zealand, Using 2‐D Full‐Waveform Inversion, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-12029, https://doi.org/10.5194/egusphere-egu2020-12029, 2020.

Cross-well 2-D seismic CT imaging method has been widely used in many fields such as oil-gas exploration and engineering geological exploration, but for the real three-dimensional space, this traditional method can only obtain the two-dimensional velocity profile between the two wells, cannot obtain the lateral geological structure outside the profile; Besides, the seismic signal received from cross-well exploration is the response of geologic body in three-dimensional space, which may be influenced by the geologic body outside the two-well profile, and that will give a result of image distortion and having an effect on geological interpretation. Based on the theory of three-dimensional acoustic wave equation, this paper implements a three-dimensional cross-well reverse-time migration imaging method to obtain the cross-well 3-D geological structure with the observed value from multiple wells by using the first-order velocity-stress acoustic wave equation and firing time imaging conditions. Calculation results of the typical theoretical models show that: The multi-well three-dimensional imaging method adopted in this paper can accurately and effectively realize the cross-well 3-D geological imaging with high resolution and reliable results. Multi-well three-dimensional imaging method can effectively obtain the cross-well three-dimensional structure distribution, which can solve the issue of hard to obtain the transverse structure change by 2-D imaging. It also can solve the imaging problems of big dip angle interface in CT imaging and obtains the true cross-well 3-D geological structure with the multiple well data, which can provide the basis for cross-well 3-D seismic exploration.

How to cite: Cheng, F. and Liu, J.: Multi-Cross-hole 3-D reverse time migration imaging, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6434, https://doi.org/10.5194/egusphere-egu2020-6434, 2020.

To find out the crustal structure and  tectonic attribute  of the Pearl river delta and offshore area(PRD), in 2015, the Guangdong Earthquake Agency collaboration with the other unit  carried out a three-dimensional joint onshore-offshore seismic detection  experiment in the PRD.  This paper processed the data of Dinghu-Gaoming-Jinwan L1 line on the west side of PDR. We utilized ray tracing and travel-time simulation  method to obtained a P-wave velocity model of the L1 profile.The study showed: Along the profile, The depth of the Moho gradually decreases from the northwestern inland 30.0km to the southwestern coastal 28.0km. Upheaval  of the Moho is between Dinghu and Gaoming. The low velocity layer in the mid-crustal  is  a heterogeneous continuum. The velocity of low velocity layer NW side is lower than the SE side, especially between Dinghu and Gaoming. The minimum velocity is 6.05km•s^-1. The deep Wuchuan-Sihui fault and Guangzhou-Enping fault  may be one of the most important channels for deep material upwelling. It is the continuum upheaval  of the Moho which from Dinghu, Gaoming on the west side of PDR to  Qingyuan, Conghua on the east side of PDR delimited by Wuchuan-Sihui fault and Guangzhou-Enping fault.

How to cite: Ye, X., Zhang, X., Lv, J., Liu, B., Wang, X., Wang, L., and Lv, Z.: Crustal Structure and tectonic attribute Revealed by a Deep Seismic Sounding Profile of Dinghu-Gaoming-Jinwan in the Pearl River Delta, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2343, https://doi.org/10.5194/egusphere-egu2020-2343, 2020.

The Pearl River estuary area, located in the middle part of the southern China coastal seismic belt, has long been considered a potential source of strong earthquakes above magnitude 7.0. To scientifically assess the potential strong earthquake risk in this area, a three-dimensional artificial seismic sounding experiment, consisting of a receiving array and seabed seismograph, was performed to reveal the deep crustal structure in this region. We used artificial ship-borne air-gun excitation shots as sources, and fixed and mobile stations as receivers to record seismic data from May to August 2015. This paper presents results along a line from the western side of the Pearl River estuary to the western side of the Baijian–Gaoming–Jinwan profile. A two-dimensional velocity structure was constructed using seismic travel-time tomography. The inversion results show that the Moho depth is 27 km in the coastal area and 30 km in the northwest of the Pearl River estuary area, indicating that the crust thins from land to sea. Two structural discontinuities and multiple low-velocity anomalies appear in the crustal section. Inside both discontinuity zones, a low-velocity layer, with a minimum velocity of 6.05kms−1, exists at a depth of about 15 km, and another, with a minimum velocity of 6.37kms−1, exists at a depth of about 21.5 km between the middle and lower crust. These low velocities suggest that the discontinuities may consist of partly molten material. Earthquakes with magnitudes higher than 5.0 occurred in the low-velocity layer along the profile. The deep Kaiping-Enping fault, rooted in the crust, may be one of the most important channels for deep material upwelling and is related to tectonic movement since the Cretaceous in the Pearl River Delta tectonic rift basin.

How to cite: Zhang, X., Ye, X., Lv, J., and Wang, X.: Crustal Structure Revealed by a Deep Seismic Sounding Profile of Baijing-Gaoming-Jinwan in the Pearl River Delta , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-3380, https://doi.org/10.5194/egusphere-egu2020-3380, 2020.

The Pearl River Delta, located in the middle of the southeast coast of south China, is a graben basin. Although this region is considered tectonically relatively inactive, many small earthquakes still occur, and multi groups of faults with different directions are well developed. To better understand the geological structures in this region, we use about 30 days of ambient noise data from 88 portable stations and 38 permanent broadband stations to obtain a high-resolution 3D upper crustal S-wave velocity model. Over 3700 Inter-station group-velocity curves were measured in the 1-10 s period range and tomographically inverted by a direct surface wave inversion method in a 0.05°×0.05°grid. The checkerboard test shows that the tomographic final resolution is 0.1°×0.1°. Our results show that in the shallow crust of the study area, the velocity distribution corresponds to surface geology and geological features. The Huizhou-Dongguan depression and the Pearl River mouth exhibit low S-wave velocity feature, while the high S-wave velocity zone corresponds to the distribution of Mesozoic granite. Some faults are almost between low velocity and high velocity zone, which may play an important role of the channel of magmatic activity. The upper crustal structure in this area is closely related to the intense magmatic tectonic activity and crustal extension since Mesozoic.

How to cite: Lyu, Z., Ye, X., Lyu, J., Zhang, X., Wang, L., and Wang, X.: Upper crustal velocity structure of Pearl River Delta, China, derived from dense-array observations of ambient noise, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7811, https://doi.org/10.5194/egusphere-egu2020-7811, 2020.

By measuring changes in radioelement concentrations, gamma-ray spectrometry is increasingly emerging as an efficient geophysical method that allows such changes to be geologically mapped according to lithology and soil type. At Maâdna crater in southern Algeria, this method has been used to monitor any changes in the composition of the target rocks that may be associated with the impact cratering process. For this purpose, several measurements were carried out in situ using a portable field gamma spectrometer. As a result, most predominantly calcareous surface lithologies, exposed on the rim and flanks of the crater, showed a very low emitted radiometric response over the three channels (K, Th, U). However, no more than 90 Cps were counted both inside and outside the crater. Such a rate is indeed expected in sedimentary rocks with low clay content, and this remains valid, as long as other exogenous mineralogical enrichments are excluded. On the other hand, the contoured radioelement concentrations maps, have demonstrated an anomalous enhanced gamma radiation levels of potassium-dominated peaks over the central part of the crater and in the surrounding wadis. Nevertheless, the central potassium anomaly is well correlated with the shallower magnetic one that has been described in previous studies (see e.g. Lamali et al., 2016). Therefore, either near the surrounding wadis or in the central part of this crater, this anomalously high level of radioactivity may be linked to an accumulation of later altered deposits. Consequently, there are no objective criteria to link these results to an impact event occurring at the Maâdna structure, similar to what was done at the Serra da Cangalha crater (Vasconcelos et al., 2012).

How to cite: Lamali, A., Hamai, L., Mokhtar, S. A., Yelles-chaouche, A., Abtout, A., Merabet, N.-E., Bentridi, S. E., Djadia, L., and Nadjemi, A.: Gamma-ray spectrometry observations to monitor a presumed meteoritic signature at Maâdna crater (Talemzane, Algeria), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2677, https://doi.org/10.5194/egusphere-egu2020-2677, 2020.

Geophysics continues to play a critical role in the future discovery of terrestrial impact structures. While the signatures within these structures may not be unique, the application of geophysics can effectively characterize them, even when they are deeply eroded or completely buried underground. In the case of Maâdna crater (33°19' N, 4°19' E), among new performed geophysical surveys, a GPR technique has been especially used to explore a supposed ejecta layer. However, GPR survey results allowed the confirmation of nonexistence of such as melting materials at Maâdna crater. Nevertheless, our different scans were interpretative against the structural context of the Maâdna structure. Indeed, most of the analyzed profiles allowed us recognizing the typical deformation effects at this structure, which can also generally be encountered at any crater-like structured site. Consequently, in view to this new resulting GPR data, even we do not definitely reject an impact origin, we are still pleading for other caratering scenarios for this structure.

How to cite: Hamai, L., Lamali, A., Yelles-chaouche, A., Abtout, A., Nadjemi, A., Merabet, N.-E., Bentridi, S.-E., Djadia, L., and mokhtar, S. A.: Georadar survey to explore a supposed ejecta layer around the Maâdna crater (Talemzane, Algeria), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2667, https://doi.org/10.5194/egusphere-egu2020-2667, 2020.

The Ivrea-Verbano Zone in the Italian Alps represents one of the most complete and best-studied cross-sections of the continental crust. Here, geological and geophysical observations indicate the presence of the Moho transition zone at shallow depth, possibly as shallow as 3 km in the location of Balmuccia in Val Sesia. Correspondingly, the Ivrea-Verbano Zone is a primary target for assembling data on the deep continental crust as well as for testing several hypotheses regarding its formation and evolution.

            Within the context of a project submitted to the International Continental Scientific Drilling Program (ICDP), the Drilling the Ivrea-Verbano zonE (DIVE) team proposes to establish three drill holes across pertinent structures within the Ivrea-Verbano Zone. Two of the planned drill holes, each with a length of ~1000 m, are within Val d’Ossola and target the Pre-Permian lower and upper section of the lower crust. The third proposed drill hole, with a length of ~4000 m, is targeting the lower most crust of the Permian magmatic system of the Ivrea-Verbano Zone in the Val Sesia, close to the Insubric Line. Combined, the three drill holes will compose a complete section of the lower crust and the Moho transition zone, and will reveal the associated structural and composition characteristics at different scales.

To bridge across the range of spatial scales and to support the drilling proposal, we have carried out active seismic surveys using an EnviroVibe source in the Val d’Ossola. These surveys combined 2D transects (in-line) with the simultaneous collection of short cross-lines, and spatially varied source points, to collect sparse 3D data with a preferential CMP coverage across strike. This survey geometry was largely controlled by environmental considerations and access for the vibrator. Accordingly, 2D profiles, both in-line and cross-line, have been processed using crooked-line geometries, which include CMPs from the 3D infill.

The very high acoustic impedance contrast of the Quaternary valley infill sediments with respect to the predominant metapelitic and gabbroic lower crustal rocks, as well as the highly attenuative nature of the sediments, were both beneficial and problematic. The former enables mapping of the valley structure, while the latter largely prevents the detection of low-amplitude reflections from within the underlying lower crustal rocks.

Here, we present the latest results of these seismic reflection surveys and discuss the observations with respect to the prevailing structure and the planning of the drilling operations. Beyond the specific objectives pursued in this study, our results have important implications with regard to the acquisition and processing of high-resolution seismic reflection data in crystalline terranes and their capacity for resolving complex, steeply dipping structures.

How to cite: Greenwood, A., Baron, L., Liu, Y., Hetényi, G., Holliger, K., Pistone, M., Zanetti, A., Ziberna, L., and Müntener, O.: Reflection seismic surveys to site the Drilling the Ivrea Verbano zonE (DIVE) proposed drill-holes, Val Sesia and Val d’Ossola, Italy., EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-18146, https://doi.org/10.5194/egusphere-egu2020-18146, 2020.

The Central and Southern Apennines are characterized by the occurrence of intense and widely spread historical and recent seismic activity, mostly located along the chain.

In this paper, we present a multi-parametric data analysis in GIS environment (Geographic Information System) with the aim of identifying and constraining the geometry (strike, dip direction and dip angle) of the seismogenic faults in areas of Central-Southern Apennines characterized by outcropping/ buried and/or active/silent faults.

We use an integrated analysis of geo-structural, seismological and gravimetric data, for the identification and geometrical description of faults with density contrast, both at the surface and at depth. At the surface, the gravity lineaments inferred by Multiscale Derivative Analysis (MDA) were compared with the Quaternary faults mapped in the study areas and with the earthquakes’ epicentral distribution. The characterization of faults at depth was instead performed by the combination of the Depth from Extreme Points (DEXP) gravity imaging method with hypocentral sections.

We tested the effectiveness of this multi-method approach at Mt. Vettore-Mt. Bove, L’Aquila basin, Mt. Massico and San Giuliano di Puglia areas (Central and Southern Apennines).

Given the effectiveness of the obtained results, this multiparametric study has been applied to other three areas of the Abruzzo-Molise region: the south-western sector of Mt. Matese, the Fucino basin and the Sulmona basin.

The Matese area was hit by a seismic sequence in 2013-2014 (M_wmax= 5.1 on December 29, 2013). Our approach showed a correlation between the epicentral distribution of the 2013-2014 Matese seismic sequence (M_w=5.0) and the MDA lineaments from gravity data. The hypocentral distibution suggests that the fault rupture does not reach the surface. Therefore, the seismogenic fault responsible of 2013-2014 Matese seismic sequence is likely a buried fault.

The Fucino basin was struck by a M_w=7.0 earthquake on January 13, 1915, causing 30,000 causalities within a large area surrounding the basin. At present, the area is characterized by scarce instrumental seismicity with low magnitude.  Our analysis highlights a good correlation between NW-SE and NE-SW well-known faults and clear gravimetric MDA maxima bordering the plain. This area can be currently considered silent but, from historical seismological studies, it is one the highest seismic risk areas of Central Apennines.

Moreover, we investigated the area of the Sulmona basin, the southwards extension of the eastern system of Central Apennines developing from Mt. Vettore, Mt. Gorzano and Mt. Gran Sasso. In historical times, the faults of the most external extensional alignment, defined as silent and considered as probable seismic gaps, activated during the 2016 Amatrice–Visso–Norcia seismic sequence. Further to the southeast, two relatively large earthquakes occurred on the eastern flank of Mt. Maiella on November 3, 1706 (M_w=6.6) and on September 26, 1933 (M_w=5.7). The Sulmona area is presently characterized by poor and low magnitude instrumental seismicity. Our multi-parametric analysis highlighted a strong correlation between MDA maxima and the Mt. Morrone normal fault bordering the western side of Mt. Maiella and the eastern side of the Sulmona basin.

How to cite: Gaudiosi, G., Paoletti, V., Nappi, R., Luiso, P., Cella, F., Florio, G., and Fedi, M.: Multiparametric data analysis for identifying active fault geometries in the Abruzzo and Molise regions (Central-Southern Appennines, Italy), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-18558, https://doi.org/10.5194/egusphere-egu2020-18558, 2020.

  The 11 March 2011 Tohoku-oki earthquake (Mw = 9.0) ruptured a 500 km-long and 200 km-wide thrust of convergent boundary between the North American and Pacific plates. The earthquake caused crustal stress field changes and triggered widespread seismic activity in the northeast Japan. The southern Fukushima area was struck by many earthquakes. The largest normal faulting (Mw = 6.6) in the area ruptured the NW-trending Yunodake fault and the NNW-trending Itozawa fault on 11 April 2011. The coseismic surface ruptures were observed along active and presumed active faults identified previously. To investigate the near-surface structure of the Itozawa fault, we conducted ground penetrating radar (GPR) profiling across the fault, and we carried out two drilling surveys in hanging and foot walls of the fault. The survey line, which length was about 50 m, was located nearby a trench site (Toda and Tsutsumi, 2013). The GPR data were collected by common-offset modes using 50, 100, and 200 MHz GPR systems (pulseEKKO PRO made by Sensors and Software Inc.), and the station spacing was 0.05 m. Furthermore, we carried out wide-angle measurements, and acquired common mid-point (CMP) ensembles at the both sides of the surface rupture to estimate the electromagnetic wave velocity used in the depth conversion of the GPR sections. The GPR sections after careful data processing show detailed structure above a depth of about 10 m. We interpreted some horizons as an event showing coseismic deformation on 11 April 2011, the past seismic event reported by Toda and Tsutsumi (2013), that informing the former event, respectively. The horizons explain accumulation of vertical displacement on the Itozawa fault.

How to cite: Kimura, H., Tsutsumi, H., Higashimaru, N., and Taniguchi, K.: Near-surface structure revealed by ground penetrating radar profiling across an inland active fault ruptured one month after the 2011 Tohoku-oki earthquake, southern Fukushima, NE Japan, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17608, https://doi.org/10.5194/egusphere-egu2020-17608, 2020.

Long-term monitoring of self-potential (SP) and electrical resistivity was conducted to examine the correlation between seismic activity and changes of these geo-electrical components in the Kyeongju area where the largest earthquake occurred in Korea. Resistivity monitoring was carried out in 2018 and 2019 but the data was not continuos occasionally because of some accidents in the field. The longest monitoring of resistivity was about 120 days and the resistivity data were acquired in 5 minutes interval. The transmitted electrical source current has 1 Hz square periodic pattern and the received voltage for the source signal was obtained in the sampling rate 10 Hz. SP data were measured in 2019 in the sampling rate of 1 kHz. The monitored resistivity and SP data are being analyzed by some graphic charts which show the variations of resistivity and SP with earthquakes of which magnitude higher than 1.0 that occurred within 4 km of the measuring site. Unfortunately, no clear correlation between the monitored geo-electrical data and seismic activity has yet been confirmed.

How to cite: Kim, H. S.: Long-term monitoring of self-potential and resistivity in Kyungju area where the largest earthquake occurred in Korea. , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-20794, https://doi.org/10.5194/egusphere-egu2020-20794, 2020.

In this study we focus on the coupled macroscopic description of the second-kind seismo-electric (SE) effect in the subsurface structure arising due to low-frequency seismic waves. Starting with the Biot poroelasticity model, we derive the equations of the coupled geomechanical-electromagnetic problem assuming mechanical excitation in the form of seismic waves (primarily Rayleigh waves), and create code for seismoelectric field simulation. We present the results of the feasibility study showing the promising possibilities for determination of non-uniform subsurface structure parameters and allowing a subsurface imaging in terms of rock elastic constants, conductivity and permeability.

The study was supported by the Russian Foundation for Basic Research (Project No. 20-05-00691).

How to cite: Alekseev, D., Gokhberg, M., Pliss, A., Goncharov, A., and Veklich, I.: Subsurface imaging with electrokinetically-induced seismoelectric signals generated by low-frequency seismic waves, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17980, https://doi.org/10.5194/egusphere-egu2020-17980, 2020.

The 2017 Ms 7.0 Sichuan Jiuzhaigou earthquake occurred at the intersection of the Tazang, Minjiang, and Huya faults on the eastern margin of the Tibetan Plateau. Since it occurred on an unmarked blind fault, it is still a controversial issue whether the fault, which triggered the earthquake, was the extension of the East Kunlun fault or the northern branch of the Huya fault. Therefore, the accurate source location is of great significance for studying the deep distribution of seismogenic faults and seismicity analysis.

We have not only collected seismic phase arrival data recorded by 24 permanent stations and 6 temporary stations, but also picked up the seismic waveform data recorded by partial permanent stations in this study. Using absolute seismic location method and relative seismic location method, we relocated the earthquake events with magnitude greater than or equal to 1.0 occurred in the Jiuzhaigou area from August to December 2017. In order to ensure reliable data quality, we selected 23422 P-wave absolute arrival times, 24734 S-wave absolute arrival times and 124519 high quality P-waveform cross correlation data of 3449 earthquake events for relocation research.

The mean value of root mean square residuals of travel time of all earthquakes decrease from 0.21s to 0.08s after relocation. The average location errors in the E-W, N-S, and vertical directions are 0.11km, 0.12km, and 0.16km, respectively. Ninety-nine percent of the earthquake events are distributed in the depth range of 1-25 km, and the dominant distribution range is 5-15 km. The result shows that the earthquakes are distributed along the strike of northwest and southeast, and the Jiuzhaigou mainshock divided these events into two clusters: northwest and southeast. From the parallel strike section, we conclude that the depth of the northwest seismic cluster is shallow with the depth range of 2-15 km, and the depth of the southeast seismic cluster is deeper with the depth range of 6-18 km. Moreover, the number of aftershocks in the northwest cluster is greater than that in the southeast cluster, but after an M 4.9 aftershock occurred in the northwest cluster on the ninety-first day after the Jiuzhaigou mainshock, the number of aftershocks in the northwest cluster began to decrease. The result provides a basis for studying the seismogenic background and seismicity of the Jiuzhaigou earthquake.

How to cite: Yu, X., Song, Q., and Deng, S.: An accurate relocation of the 2017 Ms 7.0 Sichuan Jiuzhaigou earthquake sequence and the seismicity analysis, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1802, https://doi.org/10.5194/egusphere-egu2020-1802, 2020.

Recent fast developments of satellite magnetic observations facilitate global Lithospheric Magnetic Field (LMF) modelling and their applications to subsurface tectonics. Here, the vertical component (B_z) of LMF at an altitude of 200km in Mainland China and surroundings is calculated from two global LMF models NGDC-720 and EMM2017. Next, B_z is used to invert the Curie Point Depth (CPD) by Equivalent Source Dipole (ESD) forward and Nonlinear Conjugate Gradient Method (NCGM) inversion scheme. Then, the surficial Heat Flux (HF) is derived by a simple one-dimensional steady heat conduction equation from the CPD distribution. At last, the continental seismicity is compared statistically to B_z, CPD and HF. Our essential conclusions are as follow: 1) Histograms and boxplots show that most (81.8%) earthquakes (EQs, M_s≥5.0) occurred in negative B_z areas, and more than a half (53.2%) number of EQs (corresponding to an energy percent of 94.6%) occurred inside areas with B_z between -5 and -3nT, in a period between 2004 and 2007, which is the same with the satellite data collection. When the time span is extended (most to 110 years), these phenomena maintain while weaken; 2) Most (88%) EQs occurred in areas with CPD between 10 and 30km, while only a few (7% and 5%) occurred in shallow (<10km) and deep (>30km) CPD areas, in a period between 2000 and 2010; 3) EQs seldom occurred inside cold areas (HF<50mW/m²), and are prone to occur in warm areas (HF>120mW/m²). EQs are also prone to occur along the boundaries of warm or cold areas. The mechanism of the correlations between EQs and B_z, CPD and HF maybe the lithospheric strength jumps caused by the temperature variations at boundaries between blocks with different CPDs.

How to cite: Jiao, L. and Lei, Y.: Magnetic and thermal constraints on the spatial distribution of continental seismicity, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4390, https://doi.org/10.5194/egusphere-egu2020-4390, 2020.