Upper crustal structure under Jingtai–Hezuo profile in Northeastern Tibet from topography-dependent eikonal traveltime tomography

The Northeastern Tibetan plateau records Caledonian Qilian orogeny and Cenozoic reactivation by continental collision between the Indian and Asian plates. In order to provide the constraint on the Qilian orogenic mechanism and the expansion of the plateau,wide-angle seismic data was acquired along a 430 km-long profile between Jingtai and Hezuo. There is strong height variation along the profile,which is dealt by topography flattening scheme in our crustal velocity structure reconstruction. We herein present the upper crustal P-wave velocity structure model resulting from the interpretation of first arrival dataset from topography-dependent eikonal traveltime tomography. With topography flattening scheme to process real topography along the profile,the evenness of ray coverage times of the image area（upper crust）is improved,which provides upper crustal velocity model comparable to the classic traveltime tomography（with model expansion scheme to process irregular surface）. The upper crustal velocity model shows zoning character which matcheswith the tectonic division of the Qaidam-Kunlun-West Qinling belt,the Central and Northern Qilian,and the Alax blocks along the profile. The resultant upper crustal P-wave velocity model is expected to provide important base for linkage between the mapped surface geology and deep structure or geodynamics in Northeastern Tibet.

mFAST:A MATLAB toolbox for ocean bottom seismometer refraction first-arrival traveltime tomography

First-arrival seismic traveltime tomography(FAST)is a well-established technique to estimate subsurface velocity structures.Although several existing open-source packages are available for first-arrival traveltime tomography,most were written in compiled languages and lack sufficient extendibility for new algorithms and functionalities.In this work,we develop an open-source,selfcontained FAST package based on MATLAB,one of the most popular interpreted scientific programming languages,with a focus on ocean bottom seismometer refraction traveltime tomography.Our package contains a complete traveltime tomography workflow,including ray-tracing-based first-arrival traveltime computation,linearized inversion,quality control,and high-quality visualization.We design the package as a modular toolbox,making it convenient to integrate new algorithms and functionalities as needed.At the current stage,our package is most efficient for performing FAST for two-dimensional ocean bottom seismometer surveys.We demonstrate the efficacy and accuracy of our package by using a synthetic data example based on a modified Marmousi model.

An Improved Iterative Reconstruction Algorithm for Traveltime Tomography

Traveltime tomography is a technique to reconstruct acoustic, seismic, or electromagnetic wave-speed distributions from first arrival traveltime data. The ray paths that should be used for tomographic techniques strongly depend on the wave-speed distribution. In this paper, a new method is proposed for finding out the ray paths from Fermat's principle, that means the traveltime of the ray path should be a minimum value. The problem of finding out the ray path is actually an optimum problem. Our new method uses the idea to find out the shortest path in a weighted directed graph to solve the problem. The ray paths found out by this method are used in the iterative reconstruction algorithm. Computer simulation result produced by this reconstruction algorithm is better than that by the conventional ones. It also shows that the new algorithm is effective with good convergency and stability.

Solving Traveltime Tomography with Deep Learning

This paper introduces a neural network approach for solving two-dimensional traveltime tomography(TT)problems based on the eikonal equation.The mathematical problem of TT is to recover the slowness field of a medium based on the boundary measurement of the traveltimes of waves going through the medium.This inverse map is high-dimensional and nonlinear.For the circular tomography geometry,a perturbative analysis shows that the forward map can be approximated by a vectorized convolution operator in the angular direction.Motivated by this and filtered backprojection,we propose an effective neural network architecture for the inverse map using the recently proposed BCR-Net,with weights learned from training datasets.Numerical results demonstrate the efficiency of the proposed neural networks.

First-arrival traveltime tomography for monitoring the excavation damaged zone in the Horonobe Underground Research Laboratory

Subsurface excavation results in the formation of a zone called excavation damaged zone(EDZ)around the tunnel wall.An EDZ is a major concern in the field of high-level radioactive waste disposal because it may act as a flow path after the closure of a repository.In this study,first-arrival traveltime tomography was repeatedly conducted on the EDZ at a depth of 350 m in the Horonobe Underground Research Laboratory.However,the acquired data was highly affected by the support structure on the drift wall.For proper visualization of the EDZ,information about the structure was incorporated into the inversion by modifying the model constraint.The synthetic study showed that the approach reproduced the EDZ in the model without the artifacts.The method was applied to field data,and the EDZ around the drift was detected.The inversion was extended to a time-lapse inversion to trace the changes in P-wave velocity in the EDZ.The synthetic study demonstrated that temporal changes in the P-wave velocity distribution could be detected.Data obtained from 12 surveys under open-drift conditions were analyzed by time-lapse inversion.The results indicated that the EDZ did not undergo sealing or evolution at the site for approximately seven years.

Data-Domain Wave Equation Reflection Traveltime Tomography

Estimation of an accurate macro velocity model plays an important role in seismic imag- ing and model parameter inversion. Full waveform inversion （FWI） is the classical data-domain inver- sion method. However, the misfit function of FWI is highly nonlinear, and the local optimization cannot prevent convergence of the misfit function toward local minima. To converge to the global minimum, FWI needs a good initial model or reliable low frequency component and long offset data. In this article, we present a wave-equation-based reflection traveltime tomography （WERTT） method, which can pro- vide a good background model （initial model） for FWI and （least-square） pre-stack depth migration （LS-PSDM）. First, the velocity model is decomposed into a low-wavenumber component （background velocity） and a high-wavenumber component （reflectivity）. Second, the primary reflection wave is pre- dicted by wave-equation demigration, and the reflection traveltime is calculated by an automatic picking method. Finally, the misfit function of the 12-norm of the reflection traveltime residuals is mini- mized by a gradient-based method. Numerical tests show that the proposed method can invert a good background model, which can be used as an initial model for LS-PSDM or FWI.

A Neural Network for Weighted Least－Squares Criteria of Traveltime Tomography

TH network has been applied successfully to linear optimum and some quadratic optimum problems.This paper discusses how to determine the gain function of the net amplifiers and choose appropriate parameters to solve weighted least-squares problems.To test the performance of the net,it is used for seismic traveltime tomography in computer simulation. Two media of different contrast are taken in simulation.The experimental results show that if the parameters are determined appropriately, the performance of the network is good and the results are close to the ideal ones.

Using Traveltime Tomography to Reconstruct Electrical Conductivity

A new method for reconstructing electrical conductivity distribution from electromagnetic (EM) data by using traveltime tomography is presented in this paper. Diffusive EM fields can be mathematically transformed to wavefields defined in a time like variable. The transform uniquely relates a field satisfying a diffusion equation in time, or in frequency, to an integral of the corresponding wavefield. This paper first transforms numerically calculated transient magnetic fields to wavefields. Traveltime data from a source to the receivers are estimated from the transformed wavefields. Then an iterative reconstruction algorithm is used to obtain the slowness distribution of a medium. This algorithm is an improved ART algorithm taking account of bending ray paths. The slowness distribution is transformed to electrical conductivity distribution according to their relation. The simulation result is presented in this paper.

Teleseismic P-Wave Tomography of the New Guinea-Solomon Arc System

A P-wave tomographic traveltime inversion was applied to obtain a new model of seismic velocity anomalies beneath the New Guinea-Solomon arc system(PN-SL).The P-wave traveltime data,obtained from the revised International Seismological Center catalog,were recorded by 82 seismic stations in the PN-SL.Under the constraints of the epicenter distance,magnitude,and the number of stations recorded,15009 effective P-wave traveltime data were selected from 2011 teleseismic events.The obtained model showed that the Solomon Sea Plate subducted beneath the New Britain Island along the New Britain Trench at an angle of>70°and that the slab can be traced down to a depth of approximately 800 km.Conversely,we cannot observe a high-velocity anomaly exhibited by the subducted Solomon Sea Plate in the deep mantle at the Trobriand Trench,and the slab stopped at a depth of<200 km.The double subduction of the Solomon Sea Plate strongly modified the subduction patterns of the early subducted Pacific and Australian plates in the mantle along the West Melanesian Trench and the Pocklington Trough,respectively.In addition,the subducted Solomon Sea Plate induced the melting of the upper mantle to form a low-velocity anomaly,which provided the deep dynamic source for the expansion of the Bismarck Sea.Based on the joint consideration of the tomography results and a petrological analysis,the low-velocity anomalies beneath the Solomon Sea and Woodlark Basin are closely related to the early subduction of the Pacific and Australian Plates,respectively.

基于多声源波速结构成像的岩体异常区域超前辨识方法

异常区域超前辨识对于预防地下岩土工程灾害具有重要作用。为了满足地下工程高精度探测的需求,本文提出一种层析成像方法以辨识复杂岩体结构中的异常区域,结合了走时层析、阻尼最小二乘和高斯滤波等技术。该方法克服了空洞区域辨识中速度差限制,减轻了迭代中孤立速度突变所带来的影响。我们开展了数值和室内实验量化评估最短路径法(Shortest-Path Method,SPM)、动态最短路径法(Dynamic Shortest-Path Method,DSPM)和快速扫描法(Fast Sweeping Method,FSM)等正演模拟的识别精度和计算效率。结果表明,在数值和室内实验中DSPM和FSM均能清晰地辨识出异常区域。陕西震奥矿山现场应用结果证明了该方法可利用矿山开采中爆破、微震等多类声源对矿山内部未知结构进行波速场成像。本研究不仅实现了走时层析成像方法在异常区域识别中的应用,而且为地下岩土工程中潜在风险源的探测提供了新的思路。