Xin Zhang

Bio: Xin Zhang is an academic researcher from University of Edinburgh. The author has contributed to research in topics: Inverse problem & Monte Carlo method. The author has an hindex of 8, co-authored 22 publications receiving 196 citations. Previous affiliations of Xin Zhang include University of Science and Technology of China.

3-D Monte Carlo surface wave tomography

Abstract: S U M M A R Y Seismic surface wave tomography is a tried and tested method to reveal the subsurface structure of the Earth. However, the conventional 2-step scheme of inverting first for 2-D maps of surface wave phase or group velocity and then inverting for the 3-D spatial velocity structure preserves little information about lateral spatial correlations, and introduces additional uncertainties and errors into the 3-D result. We introduce a 1-step 3-D non-linear surface wave tomography method that removes these effects by inverting for 3-D spatial structure directly from frequencydependent traveltime measurements. We achieve this using the reversible jump Markov chain Monte Carlo (McMC) algorithm with a fully 3-D model parametrization. Synthetic tests show that the method estimates the velocity model and associated uncertainties significantly better than the conventional 2-step McMC method, and that the computational cost seems to be comparable with 2-step McMC methods. The resulting uncertainties are more intuitively reasonable than those from the 2-step method, and provide directly interpretable uncertainty on volumetrics of structures of interest.

Variational full-waveform inversion

Abstract: Seismic full-waveform inversion (FWI) can produce high resolution images of the Earth's subsurface. Since full waveform modelling is significantly nonlinear with respect to velocities, Monte Carlo ...

Surface Microseismic Monitoring of Hydraulic Fracturing of a Shale-Gas Reservoir Using Short-Period and Broadband Seismic Sensors

Abstract: Online Material: Figure, table with double‐difference event locations. Hydraulic fracturing has been widely applied in development of tight sand and shale gas reservoirs, in which high‐pressure fluids are injected into target zones to enhance the reservoir permeability so that gas can be more efficiently recovered. The opening and growing of tensile fractures, as well as shearing slip along fractures during stimulation treatment are thought to be the major mechanisms inducing microseismic events around the treatment well (Shemeta and Anderson, 2010). Therefore, microseismic monitoring is a valuable approach to assess the fracturing process. For example, the locations of microseismic events are used to determine fracture network geometry, and their focal mechanisms are helpful for understanding how the fractures are stimulated. The information derived from microseismic monitoring is helpful for reservoir simulation and assessment (e.g., Rutledge and Phillips, 2003; Warpinski, 2009; Maxwell, 2010). However, injecting fluids into underground formations, especially wastewater disposal into deep wells, has caused felt or damaging earthquakes with magnitudes larger than 4 in some cases (Ellsworth, 2013). The well‐documented cases include Rocky Mountain Arsenal in the 1960s (Healy et al. , 1968), wastewater disposal in Texas (Frohlich et al. , 2011), Oklahoma (Holland, 2013), and Arkansas (Horton, 2012). In fact, an anomalous increase in earthquake activity has occurred in the central and eastern United States over the past few years, which is mainly due to deep injection of low‐pressure wastewater into deep strata or basement formations (Ellsworth, 2013). Hydraulic fracturing using high‐pressure fluids can induce a lot of earthquakes, …

Bayesian Seismic Tomography using Normalizing Flows

Abstract: We test a fully non-linear method to solve seismic tomographic problems using data consisting of observed travel times of first-arriving waves. We use variational inference to calculate the posterior probability distribution which describes the solution to the Bayesian tomographic inverse problem. The variational method is an efficient alternate to Monte Carlo methods, which seeks the best approximation to the posterior distribution. This approximation is found using an optimization framework, and the method provides fully probabilistic results. We apply a new variational method for geophysics -- normalizing flows. The method models the posterior distribution by employing a series of invertible and differentiable transforms -- the flows. By optimizing the parameters of these transforms the flows are designed to convert a simple and analytically known distribution into a good approximation of the posterior distribution. Numerical examples show that normalizing flows can provide an accurate tomographic result including full uncertainty information while significantly decreasing the computational cost compared to Monte Carlo and other variational methods. In addition, this method provides analytic solutions for the posterior distribution rather than an ensemble of posterior samples. This opens the possibility that subsequent calculations about the posterior distribution might be performed analytically.

Seismic Tomography Using Variational Inference Methods

Abstract: Seismic tomography is a methodology to image the interior of solid or fluid media, and is often used to map properties in the subsurface of the Earth. In order to better interpret the resulting images it is important to assess imaging uncertainties. Since tomography is significantly nonlinear, Monte Carlo sampling methods are often used for this purpose, but they are generally computationally intractable for large datasets and high-dimensional parameter spaces. To extend uncertainty analysis to larger systems we use variational inference methods to conduct seismic tomography. In contrast to Monte Carlo sampling, variational methods solve the Bayesian inference problem as an optimization problem, yet still provide probabilistic results. In this study, we applied two variational methods, automatic differential variational inference (ADVI) and Stein variational gradient descent (SVGD), to 2D seismic tomography problems using both synthetic and real data and we compare the results to those from two different Monte Carlo sampling methods. The results show that variational inference methods can produce accurate approximations to the results of Monte Carlo sampling methods at significantly lower computational cost, provided that gradients of parameters with respect to data can be calculated efficiently. We expect that the methods can be applied fruitfully to many other types of geophysical inverse problems.

Pattern Recognition and Machine Learning

Abstract: Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Current challenges in monitoring, discrimination, and management of induced seismicity related to underground industrial activities: A European perspective

Abstract: Due to the deep socioeconomic implications, induced seismicity is a timely and increasingly relevant topic of interest for the general public. Cases of induced seismicity have a global distribution and involve a large number of industrial operations, with many documented cases from as far back to the beginning of the twentieth century. However, the sparse and fragmented documentation available makes it difficult to have a clear picture on our understanding of the physical phenomenon and consequently in our ability to mitigate the risk associated with induced seismicity. This review presents a unified and concise summary of the still open questions related to monitoring, discrimination, and management of induced seismicity in the European context and, when possible, provides potential answers. We further discuss selected critical European cases of induced seismicity, which led to the suspension or reduction of the related industrial activities.

Fault reactivation and earthquakes with magnitudes of up to Mw4.7 induced by shale-gas hydraulic fracturing in Sichuan Basin, China

Abstract: This paper presents a timely and detailed study of significant injection-induced seismicity recently observed in the Sichuan Basin, China, where shale-gas hydraulic fracturing has been initiated and the aggressive production of shale gas is planned for the coming years. Multiple lines of evidence, including an epidemic-type aftershock sequence model, relocated hypocenters, the mechanisms of 13 large events (M W > 3.5), and numerically calculated Coulomb failure stress results, convincingly suggest that a series of earthquakes with moment magnitudes up to M W 4.7 has been induced by “short-term” (several months at a single well pad) injections for hydraulic fracturing at depths of 2.3 to 3 km. This, in turn, supports the hypothesis that they represent examples of injection-induced fault reactivation. The geologic reasons why earthquake magnitudes associated with hydraulic fracturing operations are so high in this area are discussed. Because hydraulic fracturing operations are on the rise in the Sichuan Basin, it would be beneficial for the geoscience, gas operator, regulator, and academic communities to work collectively to elucidate the local factors governing the high level of injection-induced seismicity, with the ultimate goal of ensuring that shale gas fracking can be carried out effectively and safely.