I am an associate research scholar in the Department of Geosciences, Princeton University, US. I received a PhD in geophysics from the University of Oslo and NORSAR, Norway, and an MSc in earthquake engineering from the University of Tokyo, Japan. I hold a B.E. in civil engineering and an I.E. in mechanical engineering both with the highest honor from the Tribhuvan University, Nepal. My primary research area is computational (geo)mechanics, including (an)elastic-gravitational wave propagation, postearthquake relaxation, glacial isostatic adjustment, microearthquakes; and hexahedral meshing and scientific visualization. I have a passion for teaching. I received the best lecturer award during my short stint as a lecturer of Civil Engineering immediately after finishing my Bachelor's degree.
Research Items (34)
Knowledge of deformation at plate boundaries has been improved greatly by the development of observational techniques in space geodesy. However, most theoretical and numerical models of coseismic deformation have remained very simple and do not include realistic Earth structure. Three-dimensional material heterogeneity and topography are often neglected because simple models are assumed to be sufficient and available tools cannot easily accommodate complex heterogeneity. In this study, we demonstrate the importance of three-dimensional heterogeneity using a spectral-element method that incorporates topography and 3D material properties. Using a parabolic hill model and a topographic model of the 2010 Maule earthquake, we show that topographic features can alter the shape of observed surface deformation patterns. We also estimate the coseismic surface deformation due to a model of the slip distribution of the 2015 Gorkha earthquake using realistic topography and 3D elastic structure, and find that the presence of topography causes changes in the shape of observed surface displacement patterns while material heterogeneities primarily affect the magnitude of observed displacements. Our results show that the inclusion of topography in particular can affect predictions of coseismic deformation modeling.
Although earthquake-induced gravity perturbations are frequently observed, numerical modeling of this phenomenon has remained a challenge. Due to the lack of reliable and versatile numerical tools, induced-gravity data have not been fully exploited to constrain earthquake source parameters. From a numerical perspective, the main challenge stems from the unbounded Poisson/Laplace equation that governs gravity perturbations. Additionally, the Poisson/Laplace equation must be coupled with the equation of conservation of linear momentum that governs particle displacement in the solid. Most existing methods either solve the coupled equations in a fully spherical harmonic representation, which requires models to be (nearly) spherically symmetric, or they solve the Poisson/Laplace equation in the spherical harmonics domain and the momentum equation in a discretized domain, a strategy that compromises accuracy and efficiency. We present a spectral-infinite-element approach that combines the highly accurate and efficient spectral-element method with a mapped-infinite-element method capable of mimicking an infinite domain without adding significant memory or computational costs. We solve the complete coupled momentum-gravitational equations in a fully discretized domain, enabling us to accommodate complex realistic models without compromising accuracy or efficiency. We present several coseismic and postearthquake examples and benchmark the coseismic examples against the Okubo analytical solutions. Finally, we consider gravity perturbations induced by the 1994 Northridge earthquake in a 3D model of Southern California. The examples show that our method is very accurate and efficient, and that it is stable for postearthquake simulations.
We show how the linearized equations that govern the motion of a body that undergoes deformation can be generalized to capture geometrical nonlinearities in a spectral-element formulation. Generalizing the equations adds little complexity, the main addition being that we have to track the deformation gradient. Geometrical changes due to deformation are captured using the logarithmic strain. We test the geometrically nonlinear formulation by considering numerical experiments in seismic wave propagation and cantilever beam bending and compare the results with the linearized formulation. In cases where finite deformation occurs, the effect of solving the geometrically nonlinear equations can be significant while in cases where deformation is smaller the result is similar to solving the linearized equations. We find that the time it takes to run the geometrically nonlinear simulations is on the same order of magnitude as running the linearized simulations. The limited amount of added cost and complexity suggests that we might as well solve the geometrically nonlinear equations since it does not assume anything about the size of deformations.
Accurate and efficient forward modeling methods are important for simulation of seismic wave propagation in 3D realistic Earth models and crucial for high-resolution full waveform inversion. In the presence of attenuation, wavefield simulation could be inaccurate or unstable over time if not well treated, indicating the importance of the implementation of a strong stability preserving time discretization scheme. In this study, to solve the anelastic wave equation, we choose the optimal strong stability preserving Runge-Kutta (SSPRK) method for the temporal discretization, and apply the fourth order DRP/opt MacCormack scheme for the spatial discretization. We approximate the rheological behaviors of the Earth by using the generalized Maxwell body model and use an optimization procedure to calculate the anelastic coefficients determined by the Q-law. This optimization constrains positivity of the anelastic coefficients and ensures the decay of total energy with time, resulting in a stable viscoelastic system even in the presence of strong attenuation. Moreover, we perform theoretical and numerical analyses of the SSPRK method, including the stability criteria and the numerical dispersion. Compared with the traditional fourth-order Runge-Kutta method, the SSPRK has a larger stability condition number and can better suppress numerical dispersion. We use the complex-frequency-shifted perfectly matched layer for the absorbing boundary conditions based on the auxiliary difference equation and employ the traction image method for the free-surface boundary condition on curvilinear grids representing the surface topography. Finally, we perform several numerical experiments to demonstrate the accuracy of our anelastic modeling in the presence of surface topography.
Potential field problems are ubiquitous and important in geophysics. For example, potential fields associated with gravity and magnetic anomalies are governed by the unbounded Poisson/Laplace equation. Similarly, the potential fields associated with glacial rebound, free-oscillation seismology, and earthquake-induced gravity perturbations are governed by coupled elastic-acoustic-gravitational equations. To solve these equations for complex three-dimensional (3D) models, a full geometric discretization is inevitable. However, the terms 'discretization' and 'unbounded domain' are generally incompatible. Therefore, the solution to such problems for realistic models still remains a challenge. To tackle this challenge, we developed a new method, called the spectral-infinite-element method (SIEM), by combining a mapped infinite-element approach with a spectral-element method. Spectral elements represent the domain of interest, and a single layer of infinite elements captures outer space. To solve the weak form of the Poisson/Laplace equation, we use Gauss-Legendre-Lobatto quadrature in spectral elements inside the domain of interest. Outside the domain, we use Gauss-Radau quadrature in the infinite direction and Gauss-Legendre-Lobatto quadrature in the other directions. We demonstrate the accuracy and efficiency of our method by considering several problems in geophysics, namely, gravity and magnetic anomalies, and earthquake-induced perturbations in complex 3D models. Link to the iPoster: https://agu2018fallmeeting-agu.ipostersessions.com/default.aspx?s=E2-B9-9D-6B-5B-71-49-92-0B-B2-0F-B2-69-63-A3-45
Accurate and efficient simulations of coseismic and postearthquake deformation are important for proper inferences of earthquake source parameters and subsurface structure. These simulations are often performed using a truncated halfspace model with approximate boundary conditions. The use of such boundary conditions introduces inaccuracies unless a sufficiently large model is used, which greatly increases the computational cost. To solve this problem, we develop a new approach by combining the spectral-element method with the mapped infinite-element method. In this approach, we still use a truncated model domain, but add a single outer layer of infinite elements. While the spectral elements capture the domain, the infinite elements capture the far-field boundary conditions. The additional computational cost due to the extra layer of infinite elements is insignificant. Numerical integration is performed via Gauss-Legendre-Lobatto and Gauss-Radau quadrature in the spectral and infinite elements, respectively. We implement an equivalent moment-density tensor approach and a split-node approach for the earthquake source, and discuss the advantages of each method. For postearthquake deformation, we implement a general Maxwell rheology using a second-order accurate and unconditionally stable recurrence algorithm. We benchmark our results with the Okada analytical solutions for coseismic deformation, and with the Savage & Prescott analytical solution and the PyLith finite-element code for postearthquake deformation.
Gravity anomalies induced by density heterogeneities are governed by Poisson's equation. Most existing methods for modeling such anomalies rely on its integral solution. In this approach, for each observation point, an integral over the entire density distribution needs to be carried out, and the computational cost is proportional to the number of observation points. Frequently, such methods are sensitive to high density contrasts due to inaccurate resolution of the volume integral. We introduce a new approach which directly solves a discretized form of the Poisson/Laplace equation. The main challenge in our approach involves the unbounded nature of the problem, because the potential exists in all of space. To circumvent this challenge, we combine a mapped infinite-element approach with a spectral-element method. Spectral elements represent the domain of interest, and a single layer of infinite elements captures outer space. To solve the weak form of the Poisson/Laplace equation, we use Gauss-Legendre-Lobatto quadrature in spectral elements inside the domain of interest. Outside the domain, we use Gauss-Radau quadrature in the infinite direction, and Gauss-Legendre-Lobatto quadrature in the other directions. We illustrate the efficiency and accuracy of our method by comparing calculated gravity anomalies for various density heterogeneities with corresponding analytical solutions. Finally, we consider a complex 3D model of an ore mine, which consists of both positive and negative density anomalies.
Gas extraction from the Groningen natural gas field, situated in the Netherlands, frequently induces earthquakes in the reservoir that cause damage to buildings and pose a safety hazard and a nuisance to the local population. Due to the dependence of the national heating infrastructure on Groningen gas, the short-term mitigation measures are mostly limited to a combination of spatiotemporal redistribution of gas production and strengthening measures for buildings. All options become more effective with a better understanding of both source processes and seismic wave propagation. Detailed wave propagation simulations improve both the inference of source processes from observed ground motions and the forecast of ground motions as input for hazard studies and seismic network design. The velocity structure at the Groningen site is relatively complex, including both deep high-velocity and shallow low-velocity deposits showing significant thickness variations over relatively small spatial extents. We performed a detailed three-dimensional wave propagation modelling study for an induced earthquake in the Groningen natural gas field using the spectral-element method. We considered an earthquake that nucleated along a normal fault with local magnitude of M = 3. We created a dense mesh with element size varying from 12 to 96 m, and used a source frequency of 7 Hz, such that frequencies generated during the simulation were accurately sampled up to 10 Hz. The velocity/density model is constructed using a three-dimensional geological model of the area, including both deep high-velocity salt deposits overlying the source region and shallow low-velocity sediments present in a deep but narrow tunnel valley. The results show that the three-dimensional density/velocity model in the Groningen area clearly play a large role in the wave propagation and resulting surface ground motions. The 3D structure results in significant lateral variations in site response. The high-velocity salt deposits have a dispersive effect on the radiated wavefield, reducing the seismic energy reaching the surface near the epicentre. In turn, the presence of low-velocity tunnel valley deposits can locally cause a significant increase in peak ground acceleration. Here we study induced seismicity on a local scale and use SPECFEM3D to conduct full waveform simulations and show how local velocity variations can affect seismic records.
- Jan 2018
In this study, we model acoustic waves induced by moving acoustic sources in three-dimensional (3D) underwater settings based on a spectral-element method (SEM). Numerical experiments are conducted using the SEM software package SPECFEM3D Cartesian, which facilitates fluid-solid coupling and absorbing boundary conditions. Examples presented in this article include an unbounded fluid truncated by using absorbing boundaries, and a shallow water waveguide modeled as a fluid-solid coupled system based on domain decomposition. In the numerical experiments, the SEM-computed pressures match their analytical counterparts at frequencies up to 250 Hz. SEM solutions of pressures at points behind and ahead of modeled moving acoustic sources show a frequency shift, i.e., a Doppler effect, which matches the analytical solution. This article contributes to the field of passive sonar-based detection of moving acoustic sources, and addresses the challenge of computing wave responses generated by side-scan sonar, using moving sources of continuous signals.
- Oct 2017
Offshore piling has been effective in building foundations of offshore structures, such as wind turbines, bridges, and oil rigs. Despite such merits, underwater noise due to offshore piling is considered to be its critical setback. Pile driving creates a high-level underwater sound that harms marine ecosystems. There have been studies that successfully predicted these noises by using numerical methods, for example, Finite Element Method (FEM). However, there has been no FEM study that considers anti-symmetric irregular domains and complex bathymetry to compute underwater noises in the high-frequency range (>1000 Hz) due to expensive computational costs of FEM. To bridge the gap, this work attempts to explore a novel, powerful simulation tool to efficiently obtain offshore piling noises in complex settings and in the high-frequency range. We adopted and modified an open-source large-scale parallel Spectral Element Method (SEM) wave simulator, SPECFEM3D. SEM is known to be much more efficient than FEM for wave propagation analysis problem of a very large number of elements and time steps without compromising accuracy. Our computational method can be used for prediction of offshore piling underwater noise and investigating novel piling methods, such as optimized shapes of piles or air bubbles curtains to mitigate underwater noise.
Fem1D solves the 1-D second order ordinary differential equation: a d2u/dx2 + b du/dx + c u = f, u(0) = 0, a du/dx at (x = L) = q, where both u and f are functions of x only. This package is mainly intended for educational purpose, and it demonstrates a general concept on the discretization, Gaussian quadrature, stiffness matrix, boundary conditions, matrix solution, and modular programming.
MeshAssist is a collection of tools which assists meshing of complex and realistic 2D/3D models for FEM/SPECFEM simulations. As its name suggests, it is NOT a meshing software. It is only a meshing assistant!
SPECFEM3D_GEOTECH is an open-source command-driven software for 3D slope stability analysis and simulation of 3D multistage excavation based on the spectral-element method.
Simulation of wave propagation in a microearthquake environment is often challenging due to small-scale structural and material heterogeneities. We simulate wave propagation in three different real microearthquake environments using a spectral-element method. In the first example, we compute the full wavefield in 2D and 3D models of an underground ore mine, namely the Pyhaesalmi mine in Finland. In the second example, we simulate wave propagation in a homogeneous velocity model including the actual topography of an unstable rock slope at Aaknes in western Norway. Finally, we compute the full wavefield for a weakly anisotropic cylindrical sample at laboratory scale, which was used for an acoustic emission experiment under triaxial loading. We investigate the characteristic features of wave propagation in those models and compare synthetic waveforms with observed waveforms wherever possible. We illustrate the challenges associated with the spectral-element simulation in those models.
We solve Poisson's equation by combining a spectral-element method with a mapped infinite-element method. We focus on problems in geostatics and geodynamics, where Earth's gravitational field is determined by Poisson's equation inside the Earth and Laplace's equation in the rest of space. Spectral elements are used to capture the internal field, and infinite elements are used to represent the external field. To solve the weak form of Poisson/Laplace equation, we use Gauss-Legendre-Lobatto quadrature in spectral elements inside the domain of interest. Outside the domain, we use Gauss-Radau quadrature in the infinite direction, and Gauss-Legendre-Lobatto quadrature in the other directions. We illustrate the efficiency and accuracy of the method by comparing the gravitational fields of a homogeneous sphere and the Preliminary Reference Earth Model (PREM) with (semi-)analytical solutions.
- May 2017
This work presents a new numerical approach for computingunderwater acousticwave responses due to moving underwater acoustic sources in complex underwater environments using a Spectral Element Method (SEM). The SEM is similar to the Finite Element Method(FEM), but uses a higher-order shape function with Gauss-Lobatto-Legendre quadrature, naturally creating a diagonal mass matrix. Thus, we can use fast explicit time integration, taking advantage of a diagonal mass matrix, without compromising accuracy. Therefore, the SEM is much more suitable for large-scale parallel 3D time domainwaveanalyses than the conventional FEM. In our numerical experiments, we used a large-scale parallel SEMwave simulator, SPECFEM3D. We verified the SEM solution of acoustic (fluid pressure)waves in a 3D acoustic fluid setting of an infinite extent, induced by a moving point source, by using its analytical counterpart. Numerical experiments showed that our tool accurately accommodates wave behavior at fluid-solid interfaces of complex geometries and infinite extents of water and solids, truncated using absorbing boundary conditions. Due to such versatility, our tool can be used for forward and inverse acoustic waveanalyses in any complex underwater systems of large extents (e.g., shallow water and deep ocean).
The subduction earthquake cycle includes multiple stages of deformation that are caused by the evolution of stress in the crust and upper mantle. In recent years, observations of deformation at plate boundaries have been greatly improved by the development of techniques in space geodesy, including GPS, InSAR, and GRACE. However, models of seismic deformation remain limited, and are unable to account for 3D structure in topography and material properties. We seek to investigate the importance of 3D structure using a spectral-element method that incorporates complex fault geometry, topography, and heterogeneous material properties in a (non)linear viscoelastic domain. The 2010 Maule, Chile earthquake is used as a case study to examine whether 3D structure can affect the predictions of seismic deformation models.
We incorporate 3D anelastic attenuation into the collocated-grid finite-difference method on curvilinear grids (Zhang et al., 2012), using the rheological model of the generalized Maxwell body (Emmerich and Korn, 1987; Moczo and Kristek, 2005; Käser et al., 2007). We follow a conventional procedure to calculate the anelastic coefficients (Emmerich and Korn, 1987) determined by the Q(ω)-law, with a modification in the choice of frequency band and thus the relaxation frequencies that equidistantly cover the logarithmic frequency range. We show that such an optimization of anelastic coefficients is more accurate when using a fixed number of relaxation mechanisms to fit the frequency independent Q-factors. We use curvilinear grids to represent the surface topography. The velocity-stress form of the 3D isotropic anelastic wave equation is solved with a collocated-grid finite-difference method. Compared with the elastic case, we need to solve additional material-independent anelastic functions (Kristek and Moczo, 2003) for the mechanisms at each relaxation frequency. Based on the stress-strain relation, we calculate the spatial partial derivatives of the anelastic functions indirectly thereby saving computational storage and improving computational efficiency. The complex-frequency-shifted perfectly matched layer (CFS-PML) is used for the absorbing boundary condition based on the auxiliary difference equation (Zhang and Shen, 2010). The traction image method (Zhang and Chen, 2006) is employed for the free-surface boundary condition. We perform several numerical experiments including homogeneous full-space models and layered half-space models, considering both flat and 3D Gaussian-shape hill surfaces. The results match very well with those of the spectral-element method (Komatitisch and Tromp, 2002; Savage et al., 2010), verifying the simulations by our method in the anelastic model with surface topography.
- Dec 2015
- American Geophysical Union, Fall Meeting 2013
Gravitational perturbations induced by particle motions are governed by the Poisson/Laplace equation, whose domain includes all of space. Due to its unbounded nature, obtaining an accurate numerical solution is very challenging. Consequently, gravitational perturbations are generally ignored in simulations of global seismic wave propagation, and only the unperturbed equilibrium gravitational field is taken into account. This so-calledCowling approximation is justified for relatively short-period waves (periods less than 250 s), but is invalid for free-oscillation seismology. Existing methods are usually based on spherical harmonic expansions. Most methods are either limited to spherically symmetric models or have to rely on costly iterative implementation procedures. We propose a spectral-infinite-element method to solve wave propagation in a self-gravitating Earth model. The spectral-infinite-element method combines the spectral-element method with the infinite-element method. Spectral elements are used to capture the internal field, and infinite elements are used to represent the external field. To solve the weak form of the Poisson/Laplace equation, we employ Gauss-Legendre-Lobattoquadrature in spectral elements. In infinite elements, Gauss-Radau quadrature is used in the radial direction whereas Gauss-Legendre-Lobatto quadrature is used in the lateral directions. Infinite elements naturally integrate with spectral elements, thereby avoiding an iterative implementation. We demonstrate the accuracy of the method by comparing our results with a spherical harmonics method. The new method empowers us to tackle several problems in long-period seismology accurately and efficiently.
- Dec 2013
- American Geophysical Union, Fall Meeting 2013
The complete set of governing equations for global dynamic and quasistatic problems --such as post-seismic and post-glacial rebound, tidal loading, and long-period seismology-- involves a coupling between the conservation laws of continuum mechanics and Poisson/Laplace's equation. For dynamic problems, such as seismic wave propagation and the free oscillations of the Earth, it is possible to decouple Poisson's equation using an explicit time marching scheme so that it can be solved independently. For quasistatic problems, such as glacial isostatic adjustment and tidal loading, inertia is neglected, requiring an implicit time marching scheme. In the latter case, Poisson's equation cannot be decoupled. Although an explicit time scheme with an independent Poisson's solver is generally fast, such an approach is limited by conditional stability, such that a very large number of time steps is often necessary. On the other hand, an implicit time scheme coupled with Poisson's equation is generally slow but unconditionally stable. In both cases, the unbounded and large-scale nature of the problem poses numerical challenges, particularly for 3D Earth models. Most of the existing methods use spherical harmonics to solve the unbounded Poisson/Laplace's equation. Such methods are often limited to spherically-symmetric models or have to rely on iterative procedures. In view of these challenges, we develop a parallel software package based on the spectral-element method combined with a mapped infinite-element approach. While the spectral-element method is used within the Earth model, the infinite-element approach is employed in the outer region. In the infinite element approach, a so-called infinite-element layer is used to mimic all of space. The outermost edges of an element in the infinite-element layer are mapped to infinity in order to reproduce the behavior of gravitational potential outside the domain of interest, such that the potential decays to zero at infinity. Gauss-Lobatto-Legendre (GLL) quadrature is used for numerical integration in spectral elements. Since GLL quadrature cannot be used in infinite elements due to a singularity, we use Gauss-Radau quadrature instead. Spectral and infinite elements share identical quadrature points on infinite-element boundaries, thereby providing a natural coupling of the infinite-element method with the spectral-element method. We use a generalized Maxwell rheology for viscoelastic deformation and accommodate topography and ellipticity. Both explicit and implicit time schemes are implemented in order to address a range of problems, including long-period seismology, glacial rebound, tidal loading, etc.
- Jul 2012
We implement a spectral-element method for 3D time-independent elastoplastic problems in geomechanics. As a first application, we use the method for slope stability analyses ranging from small to large scales. The implementation employs an element-by-element preconditioned conjugate-gradient solver for efficient storage. The program accommodates material heterogeneity and complex topography. Either simple or complex water table profiles may be used to assess effects of hydrostatic pressure. Both surface loading and pseudostatic seismic loading are implemented. For elastoplastic behavior of slopes to be simulated, a Mohr-Coulomb yield criterion is employed using an initial strain method (i.e., a viscoplastic algorithm). For large-scale problems, the software is parallelized on the basis of domain decomposition using Message Passing Interface. Strong-scaling measurements demonstrate that the parallelized software performs efficiently. We validate our spectral-element results against several other methods and apply the technique to simulate failure of an earthen embankment and a mountain slope.
- Jun 2012
We implement a 3D spectral-element method for multistage excavation problems. To simulate excavation in elastoplastic soils, we employ a Mohr–Coulomb yield criterion using an initial strain method. We parallelize the software based on non-overlapping domain decomposition using MPI. We verify the uniqueness principle for multistage excavation in linear elastic materials. We validate our serial and parallel programs, and illustrate several examples of multistage excavation in elastoplastic materials. Finally, we apply our software to a model of the Pyhäsalmi ore mine in Finland. Strong-scaling performance tests involving multistage excavation show that the parallel program performs reasonably well for large-scale problems.
- Mar 2012
yen is considered physically open, and, therefore, will likely experience drift of the injected CO2 towards the Northeast, through gradual mixing and expulsion of saline groundwater. This offers a unique opportunity for studying the behavior of CO2 in subsurface saline aquifers. Six slim-hole wells have been drilled so far and several new monitoring wells are planned in what becomes a “well park”. In this study, we try to use induced seismicity to monitor the injection fluid in the test site. A precise estimation of the location and magnitude of the microearthquake will be important to investigate the link between the injection and the sudden stress release as a microearthquake. In addition, the spatial and temporal patterns of the occurrences of microearthquakes might help us better constrain the migration pattern of the injected fluid.
We use time-reversal imaging for microearthquake location and investigate the possibility of a simultaneous qualitative moment-tensor estimation. We cross-correlate the data with the synthetic strain Green's tensor and stack individually for each moment-tensor component. The objective function for the source location is then formulated as the squared sum of those stacked components. The maximum value of the objective function corresponds to the estimated source location and origin time. Similarly, the corresponding stacked components at the estimated source location give the entire time history of the qualitative estimation of the moment tensor. We apply the method to synthetic data of various types of moment-tensor sources computed for a complex and heterogeneous model, namely the Pyhäsalmi ore mine in Finland. We also test the method with the same data adding white noise up to 40% of the absolute maximum. Although the method is computationally intensive, it is fully automatic and can easily be adapted to parallel processing. The preliminary results show that the method is robust and reliable.
Most earthquake location methods require phase identification and arrival-time measurements. These methods are generally fast and efficient but not always applicable to microearthquake data with low signal-to-noise ratios because the phase identification might be very difficult. The migration-based source location methods, which do not require an explicit phase identification, are often more suitable for such noisy data. Whereas some existing migration-based methods are computationally intensive, others are limited to a certain type of data or make use of only a particular phase of the signal. We have developed a migration-based source location method especially applicable to data with relatively low signal-to-noise ratios. We projected seismograms onto the ray coordinate system for each potential source-receiver configuration and subsequently computed their envelopes. The envelopes were time shifted according to synthetic P- and S-wave arrival times (computed using an eikonal solver) and stacked for a
- Jan 2010
- SEG Technical Program Expanded Abstracts 2010
Monitoring of microearthquakes is routinely conducted in various environments such as hydrocarbon and geothermal reservoirs, mines, dams, seismically active faults, volcanoes, nuclear power plants and CO2 storages. In many of these cases the handled data is sensitive and the interpretation of the data may be vital. In some cases, such as during mining or hydraulic fracturing activities, the number of microearthquakes is very large with tens to thousands of events per hour. In others, almost no events occur during a week and furthermore, it might not be anticipated that many events occur at all. However, the general setup of seismic networks, including surface and downhole stations, is usually optimized to record as many microearthquakes as possible, thereby trying to lower the detection threshold of the network. This process is obviously limited to some extent. Most microearthquake location techniques take advantage of a combination of P- and S-wave onset times that often can be picked reliably in an automatic mode. Moreover, when using seismic wave onset times, sometimes in combination with seismic wave polarization, these methods are more accurate compared to migration-based location routines. However, many events cannot be located because their magnitude is too small, i.e. the P- and/or S-wave onset times cannot be picked accurately on a sufficient number of receivers. Nevertheless, these small events are important for the interpretation of the processes that are monitored and even an inferior estimate of event locations and strengths is valuable information. Moreover, the smaller the event the more often such events statistically occur and the more important such additional information becomes. In this study we try to enhance the performance of any microseismic network, providing additional estimates of event locations below the actual detection threshold. We present a migration-based event location method, where we project the recorded seismograms onto the ray coordinate system, which corresponds to a configuration of trial sources and the real receiver network. A time window of predefined length is centered on the arrival time of the related phase that is calculated for the same grid of trial locations. The area spanned by the time window below the computed envelope is stacked for each component (L, T, Q) individually. Subsequently, the objective function is formulated as the squared sum of the stacked values. To obtain the final location, we apply a robust global optimization routine called differential evolution, which provides the maximum value of the objective function. This method provides a complete algorithm with a minimum of control parameters making it suitable for automated processing. The method can be applied to both single and multi-component data, and either P or S or both phases can be used. As a result, this method allows for a flexible application to a wide range of data. Synthetic data were computed for a complex and heterogeneous model of an ore mine and we applied this method to real, observed microearthquake data.
Most earthquake location methods require phase identification and arrival time measurements. These methods are generally fast and efficient, but not always applicable to microearthquake data with low signal-to-noise ratios, where the phase identification might be very difficult. Migration-based source location methods, which do not require the explicit phase identification are often more suitable for such low quality data. While some of the existing migration-based methods are computationally intensive, others are limited to a certain type of data or make use of only a particular phase of the signal. We present a migration-based event location method especially applicable to data with very low signal-to-noise ratios. In this method, we project the seismograms onto the local ray-centered coordinate system (LTQ) for each source-receiver configuration and compute their envelopes. A time window of predefined length is centered on the computed arrival time of the corresponding phase in each seismogram. We stack the envelopes for these time windows for each component (L, T, Q) individually. Subsequently, the objective function is formulated as the squared sum of the stacked values. We apply a robust global optimization routine called differential evolution to identify the source location. Our procedure provides a complete algorithm with a minimum of control parameters making it suitable for automated processing. The method can be applied to both single and multi-component data, and either P or S or both phases can be used. As a result, this method allows for a flexible application to a wide range of data. We performed an extensive test using synthetic microearthquake data superimposed with 30% white noise. These synthetic data were computed for a complex and heterogeneous model of the Pyhäsalmi ore mine in Finland. Finally, we applied this method to observed microearthquake data.
- Jun 2009
- 71st EAGE Conference and Exhibition incorporating SPE EUROPEC 2009
Microseismic monitoring in mines is a well-established tool used to minimise risk for personnel and to optimise production. Standard monitoring focuses on event locations, but information on detailed source parameters becomes more important. The estimatio
- Mar 2009
- EAGE Passive Seismic Workshop - Exploration and Monitoring Applications 2009
Microseismic monitoring of hydraulic fractures is an important tool for imaging fracture networks and optimizing the reservoir engineering of the stimulation. The range of magnitudes of the recorded microseisms depends at the lower limit on the array sensitivity; while the upper limit varies significantly from site to site. In this paper the variation in the microseismic magnitude range is examined and compared with the injection and site characteristics. Although there are numerous potential factors effecting the seismic deformation, the energy of the pumping and state of stress appear to be the two dominant factors. However, interaction with pre-existing faults also results in increased deformation. In this paper these factors are examined using the seismic injection efficiency, defined as the ratio of seismic energy release to the hydraulic energy expended during the injection. Ultimately, this can potentially be used to design the stimulation to maximize the deformation. Characterization of the seismogenic potential is also important for seismic hazard assessment, as well as the design of passive monitoring.
- Jan 2009
Standard microseismic monitoring nowadays focuses on event locations, but information on detailed source parameters is rarely extracted The estimation of full moment tensor solutions of microearthquakes may provide valuable information on presence and geometry of faults as well as stress fields and changes thereof Therefore, we included a full moment tensor inversion using automatically picked P-wave first motion amplitudes m addition to the computations of source spectra and related source parameters We tested the routine using synthetic polarities and synthetic waveforms computed in a complex 3-D velocity model In a first application, we investigated mining-induced seismicity m the Pyhasalmi ore mine, Finland Approximately 50% of the events show strong non-double couple components Known explosions could be retrieved, being used as quality control of the method.