Summary Induced Polarization (IP) effects can significantly affect and superimpose the inductive earth response, leading to heavily distorted data and, if overlooked, false geological interpretation. In this paper, we implemented the Levenberg-Marquardt (LM) and very fast simulated annealing (VFSA) algorithms to recover induced polarization effects from central loop transient electromagnetic (TEM) data. To incorporate the IP effect in the TEM response, we used the Cole-Cole parameterization, maximum phase angle (MPA), maximum imaginary conductivity (MIC), and Jeffrey transform of Cole-Cole parameters. The result of 1D forward calculation and inversion of synthetic TEM data revealed that the Cole-Cole parametrization is more robust and reliable than MPA, MIC, and Jeffrey transform, and that the synthetic data were well fitted and IP parameters well recovered using this model. However, the incorporation of the IP effect leads to a highly non-linear and non-unique inverse problem which requires an accurate starting model, especially for LM inversion. To evaluate the performance of our algorithm using field data, we carried out a 1D inversion of TEM data acquired along a profile that traverses a waste site located near Cologne, Germany. Furthermore, to obtain a priori information and validate the result of TEM data modeling, we conducted an electrical resistivity tomography (ERT) and time-domain IP (TDIP) survey along the TEM profile. A 2D inversion was used to retrieve the Cole-Cole parameters as input for TEM interpretation. By including the IP information, the TEM field data can be explained quantitively, and a consistent and improved interpretation of the waste body is achieved.
摘要 感应极化（IP）效应会严重影响和叠加感应地球响应，导致数据严重失真，如果被忽视，还会造成错误的地质解释。在本文中，我们采用 Levenberg-Marquardt (LM) 算法和快速模拟退火 (VFSA) 算法，从中心环瞬态电磁 (TEM) 数据中恢复感应极化效应。为了将 IP 效应纳入 TEM 响应，我们使用了科尔-科尔参数化、最大相位角 (MPA)、最大虚电导率 (MIC) 和科尔-科尔参数的杰弗里变换。合成 TEM 数据的一维正演计算和反演结果表明，Cole-Cole 参数化比 MPA、MIC 和 Jeffrey 变换更加稳健可靠，而且使用该模型可以很好地拟合合成数据并恢复 IP 参数。然而，IP效应的加入导致了一个高度非线性和非唯一的逆问题，这就需要一个精确的起始模型，特别是对于 LM 反演。为了利用现场数据评估我们算法的性能，我们对沿着穿越德国科隆附近一个垃圾场的剖面获取的 TEM 数据进行了一维反演。此外，为了获得先验信息并验证 TEM 数据建模的结果，我们沿 TEM 剖面进行了电阻率层析成像 (ERT) 和时域 IP (TDIP) 勘测。二维反演用于检索 Cole-Cole 参数，作为 TEM 解释的输入。通过加入 IP 信息，可以对 TEM 现场数据进行量化解释，并实现对废料体的一致和更好的解释。
{"title":"Recovering Induced Polarization Effects from 1-D Coupled Inversion of Transient Electromagnetic Data","authors":"Fereydoun Sharifi, Bülent Tezkan, Ismael M Ibraheem, Rainer Bergers, Pritam Yogeshwar","doi":"10.1093/gji/ggae237","DOIUrl":"https://doi.org/10.1093/gji/ggae237","url":null,"abstract":"Summary Induced Polarization (IP) effects can significantly affect and superimpose the inductive earth response, leading to heavily distorted data and, if overlooked, false geological interpretation. In this paper, we implemented the Levenberg-Marquardt (LM) and very fast simulated annealing (VFSA) algorithms to recover induced polarization effects from central loop transient electromagnetic (TEM) data. To incorporate the IP effect in the TEM response, we used the Cole-Cole parameterization, maximum phase angle (MPA), maximum imaginary conductivity (MIC), and Jeffrey transform of Cole-Cole parameters. The result of 1D forward calculation and inversion of synthetic TEM data revealed that the Cole-Cole parametrization is more robust and reliable than MPA, MIC, and Jeffrey transform, and that the synthetic data were well fitted and IP parameters well recovered using this model. However, the incorporation of the IP effect leads to a highly non-linear and non-unique inverse problem which requires an accurate starting model, especially for LM inversion. To evaluate the performance of our algorithm using field data, we carried out a 1D inversion of TEM data acquired along a profile that traverses a waste site located near Cologne, Germany. Furthermore, to obtain a priori information and validate the result of TEM data modeling, we conducted an electrical resistivity tomography (ERT) and time-domain IP (TDIP) survey along the TEM profile. A 2D inversion was used to retrieve the Cole-Cole parameters as input for TEM interpretation. By including the IP information, the TEM field data can be explained quantitively, and a consistent and improved interpretation of the waste body is achieved.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"39 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141584775","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary The ratio of the magnetic power spectrum and the secular variation spectrum measured at the Earth’s surface provides a time scale τsv(l) as a function of spherical harmonic degree l. τsv is often assumed to be representative of time scales related to the dynamo inside the outer core and its scaling with l is debated. To assess the validity of this surmise and to study the time variation of the geomagnetic field $dot{B}$ inside the outer core, we introduce a magnetic time-scale spectrum τ(l, r) that is valid for all radius r above the inner core and reduces to the usual τsv at and above the core–mantle boundary (CMB). We study τ in a numerical geodynamo model. At the CMB, we find that τ ∼ l−1 is valid at both the large and small scales, in agreement with previous numerical studies on τsv. Just below the CMB, the scaling undergo a sharp transition at small l. Consequently, in the interior of the outer core, τ exhibits different scaling at the large and small scales, specifically, the scaling of τ becomes shallower than l−1 at small l. We find that this transition at the large scales stems from the fact that the horizontal components of the magnetic field evolve faster than the radial component in the interior. In contrast, the magnetic field at the CMB must match onto a potential field, hence the dynamics of the radial and horizontal magnetic fields are tied together. The upshot is τsv becomes unreliable in estimating time scales inside the outer core. Another question concerning τ is whether an argument based on the frozen-flux hypothesis can be used to explain its scaling. To investigate this, we analyse the induction equation in the spectral space. We find that away from both boundaries, the magnetic diffusion term is negligible in the power spectrum of $dot{B}$. However, $dot{B}$ is controlled by the radial derivative in the induction term, thus invalidating the frozen-flux argument. Near the CMB, magnetic diffusion starts to affect $dot{B}$ rendering the frozen-flux hypothesis inapplicable. We also examine the effects of different velocity boundary conditions and find that the above results apply for both no-slip and stress-free conditions at the CMB.
{"title":"Scaling of the geomagnetic secular variation time scale","authors":"Yue-Kin Tsang, Chris A Jones","doi":"10.1093/gji/ggae234","DOIUrl":"https://doi.org/10.1093/gji/ggae234","url":null,"abstract":"Summary The ratio of the magnetic power spectrum and the secular variation spectrum measured at the Earth’s surface provides a time scale τsv(l) as a function of spherical harmonic degree l. τsv is often assumed to be representative of time scales related to the dynamo inside the outer core and its scaling with l is debated. To assess the validity of this surmise and to study the time variation of the geomagnetic field $dot{B}$ inside the outer core, we introduce a magnetic time-scale spectrum τ(l, r) that is valid for all radius r above the inner core and reduces to the usual τsv at and above the core–mantle boundary (CMB). We study τ in a numerical geodynamo model. At the CMB, we find that τ ∼ l−1 is valid at both the large and small scales, in agreement with previous numerical studies on τsv. Just below the CMB, the scaling undergo a sharp transition at small l. Consequently, in the interior of the outer core, τ exhibits different scaling at the large and small scales, specifically, the scaling of τ becomes shallower than l−1 at small l. We find that this transition at the large scales stems from the fact that the horizontal components of the magnetic field evolve faster than the radial component in the interior. In contrast, the magnetic field at the CMB must match onto a potential field, hence the dynamics of the radial and horizontal magnetic fields are tied together. The upshot is τsv becomes unreliable in estimating time scales inside the outer core. Another question concerning τ is whether an argument based on the frozen-flux hypothesis can be used to explain its scaling. To investigate this, we analyse the induction equation in the spectral space. We find that away from both boundaries, the magnetic diffusion term is negligible in the power spectrum of $dot{B}$. However, $dot{B}$ is controlled by the radial derivative in the induction term, thus invalidating the frozen-flux argument. Near the CMB, magnetic diffusion starts to affect $dot{B}$ rendering the frozen-flux hypothesis inapplicable. We also examine the effects of different velocity boundary conditions and find that the above results apply for both no-slip and stress-free conditions at the CMB.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"10 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141568124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary Large-scale and high-resolution seismic modelling are very significant to simulating seismic waves, evaluating earthquake hazards, and advancing exploration seismology. However, achieving high-resolution seismic modelling requires substantial computing and storage resources, resulting in a considerable computational cost. To enhance computational efficiency and performance, recent heterogeneous computing platforms, such as Nvidia Graphics Processing Units (GPUs), natively support half-precision floating-point numbers (FP16). FP16 operations can privide faster calculation speed, lower storage requirements and greater performance enhancement over single-precision floating-point numbers (FP32), thus providing significant benefits for seismic modelling. Nevertheless, the inherent limitation of fewer 16-bit representations in FP16 may lead to severe numerical overflow, underflow, and floating-point errors during computation. In this study, to ensure stable wave equation solutions and minimize the floating-point errors, we employ a scaling strategy to adjust the computation of FP16 arithmetic operations. For optimal GPU floating-point performance, we implement a 2-way single instruction multiple data (SIMD) within the floating-point units (FPUs) of CUDA cores. Moreover, we implement an earthquake simulation solver for FP16 operations based on curvilinear grid finite-difference method (CGFDM) and perform several earthquake simulations. Comparing the results of wavefield data with the standard CGFDM using FP32, the errors introduced by FP16 are minimal, demonstrating excellent consistency with the FP32 results. Performance analysis indicates that FP16 seismic modelling exhibits a remarkable improvement in computational efficiency, achieving a speedup of approximately 1.75 and reducing memory usage by half compared to the FP32 version.
{"title":"Enhancing computational efficiency in 3-D seismic modelling with half-precision floating-point numbers based on the curvilinear grid finite-difference method","authors":"Jialiang Wan, Wenqiang Wang, Zhenguo Zhang","doi":"10.1093/gji/ggae235","DOIUrl":"https://doi.org/10.1093/gji/ggae235","url":null,"abstract":"Summary Large-scale and high-resolution seismic modelling are very significant to simulating seismic waves, evaluating earthquake hazards, and advancing exploration seismology. However, achieving high-resolution seismic modelling requires substantial computing and storage resources, resulting in a considerable computational cost. To enhance computational efficiency and performance, recent heterogeneous computing platforms, such as Nvidia Graphics Processing Units (GPUs), natively support half-precision floating-point numbers (FP16). FP16 operations can privide faster calculation speed, lower storage requirements and greater performance enhancement over single-precision floating-point numbers (FP32), thus providing significant benefits for seismic modelling. Nevertheless, the inherent limitation of fewer 16-bit representations in FP16 may lead to severe numerical overflow, underflow, and floating-point errors during computation. In this study, to ensure stable wave equation solutions and minimize the floating-point errors, we employ a scaling strategy to adjust the computation of FP16 arithmetic operations. For optimal GPU floating-point performance, we implement a 2-way single instruction multiple data (SIMD) within the floating-point units (FPUs) of CUDA cores. Moreover, we implement an earthquake simulation solver for FP16 operations based on curvilinear grid finite-difference method (CGFDM) and perform several earthquake simulations. Comparing the results of wavefield data with the standard CGFDM using FP32, the errors introduced by FP16 are minimal, demonstrating excellent consistency with the FP32 results. Performance analysis indicates that FP16 seismic modelling exhibits a remarkable improvement in computational efficiency, achieving a speedup of approximately 1.75 and reducing memory usage by half compared to the FP32 version.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"20 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141568172","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary For seismographic stations with short acquisition duration, the signal-to-noise ratios (SNRs) of ambient noise cross-correlation functions (CCFs) are typically low, preventing us from accurately measuring surface wave dispersion curves or waveform characteristics. In addition, with noisy CCFs, it is difficult to extract relatively weak signals such as body waves. In this study, we propose to use local attributes to improve the SNRs of ambient noise CCFs, which allows us to enhance the quality of CCFs for stations with limited acquisition duration. Two local attributes: local cross-correlation and local similarity, are used in this study. The local cross-correlation allows us to extend the dimensionality of daily CCFs with computational costs similar to global cross-correlation. Taking advantage of this extended dimensionality, the local similarity is then used to measure non-stationary similarity between the extended daily CCFs with a reference trace, which enables us to design better stacking weights to enhance coherent features and attenuate incoherent background noises. Ambient noise recorded by several broadband stations from the USArray in North Texas and Oklahoma, the Superior Province Rifting EarthScope Experiment in Minnesota and Wisconsin and a high-frequency nodal array deployed in the northern Los Angeles basin are used to demonstrate the performance of the proposed approach for improving the SNR of CCFs.
{"title":"Improving signal-to-noise ratios of ambient noise cross-correlation functions using local attributes","authors":"Bin He, Hejun Zhu, David Lumley","doi":"10.1093/gji/ggae228","DOIUrl":"https://doi.org/10.1093/gji/ggae228","url":null,"abstract":"Summary For seismographic stations with short acquisition duration, the signal-to-noise ratios (SNRs) of ambient noise cross-correlation functions (CCFs) are typically low, preventing us from accurately measuring surface wave dispersion curves or waveform characteristics. In addition, with noisy CCFs, it is difficult to extract relatively weak signals such as body waves. In this study, we propose to use local attributes to improve the SNRs of ambient noise CCFs, which allows us to enhance the quality of CCFs for stations with limited acquisition duration. Two local attributes: local cross-correlation and local similarity, are used in this study. The local cross-correlation allows us to extend the dimensionality of daily CCFs with computational costs similar to global cross-correlation. Taking advantage of this extended dimensionality, the local similarity is then used to measure non-stationary similarity between the extended daily CCFs with a reference trace, which enables us to design better stacking weights to enhance coherent features and attenuate incoherent background noises. Ambient noise recorded by several broadband stations from the USArray in North Texas and Oklahoma, the Superior Province Rifting EarthScope Experiment in Minnesota and Wisconsin and a high-frequency nodal array deployed in the northern Los Angeles basin are used to demonstrate the performance of the proposed approach for improving the SNR of CCFs.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"33 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141546896","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary Using forward mantle convection models starting at 140 Ma, and assimilating plate reconstructions as surface velocity boundary condition, we predict present-day mantle structure and compare them with tomography models, using geoid as an additional constraint. We explore a wide model parameter space, such as different values of Clapeyron slope and density change across 660 km, density and viscosity of the thermochemical piles at the core-mantle boundary (CMB), internal heat generation rate, and model initiation age. We also investigate the effects of different strengths of a weak layer below 660 km and weaker asthenosphere and slabs. Our results suggest that slab structures at different subduction zones are sensitive to the viscosity of the asthenosphere, strength of slabs, values of Clapeyron slope and the density and viscosity of the thermochemical piles, while different internal heat generation rates do not affect the slab structures. We find that with a moderately weak asthenosphere (1020 Pas) and strong slabs, the predicted slab structures are consistent with the tomography models, and the observed geoid is also matched well. Moreover, our models successfully reproduce the degree-2 structure of the lower mantle beneath Africa and the Pacific, also known as Large Low Shear Velocity provinces (LLSVPs). A moderate Clapeyron slope of -2.5 MPa/K at 660 km aids in slab stagnation while higher values result in massive slab accumulation at that depth, ultimately leading to slab avalanches. We also find that the convective patterns in the thermal and thermochemical cases with slightly denser LLSVPs are similar, although the geoid amplitudes are lower for the latter. However, with more dense LLSVPs, the slabs cannot perturb them and no plumes are generated. Plumes arise as thermal instabilities from the edges of the LLSVPs, when cold and viscous slabs perturb them. While our predicted plume locations are consistent with the observed hotspot locations, matching the plume structures in tomography models is difficult. These plumes are essential in fitting the finer features of the observed geoid. In longer-duration models, more voluminous subducted material reaches the CMB, which tends to erode the LLSVPs significantly, and yields a poor fit to the observed geoid. Our results suggest that with the presence of a thin, moderately weak layer below 660 km, a slightly dense LLSVP, and Clapeyron slope of -2.5 MPa/K, the velocity anomalies in seismic tomography and the long-wavelength geoid can be matched well. One of the limitations of our models is that the assimilated plate motion history may be too short to overcome arbitrary initial conditions effects. Also, assimilated true plate velocities in our models may not represent the true convective vigor of the Earth.
{"title":"Present day mantle structure from global mantle convection models since the Cretaceous","authors":"Debanjan Pal, Attreyee Ghosh","doi":"10.1093/gji/ggae231","DOIUrl":"https://doi.org/10.1093/gji/ggae231","url":null,"abstract":"Summary Using forward mantle convection models starting at 140 Ma, and assimilating plate reconstructions as surface velocity boundary condition, we predict present-day mantle structure and compare them with tomography models, using geoid as an additional constraint. We explore a wide model parameter space, such as different values of Clapeyron slope and density change across 660 km, density and viscosity of the thermochemical piles at the core-mantle boundary (CMB), internal heat generation rate, and model initiation age. We also investigate the effects of different strengths of a weak layer below 660 km and weaker asthenosphere and slabs. Our results suggest that slab structures at different subduction zones are sensitive to the viscosity of the asthenosphere, strength of slabs, values of Clapeyron slope and the density and viscosity of the thermochemical piles, while different internal heat generation rates do not affect the slab structures. We find that with a moderately weak asthenosphere (1020 Pas) and strong slabs, the predicted slab structures are consistent with the tomography models, and the observed geoid is also matched well. Moreover, our models successfully reproduce the degree-2 structure of the lower mantle beneath Africa and the Pacific, also known as Large Low Shear Velocity provinces (LLSVPs). A moderate Clapeyron slope of -2.5 MPa/K at 660 km aids in slab stagnation while higher values result in massive slab accumulation at that depth, ultimately leading to slab avalanches. We also find that the convective patterns in the thermal and thermochemical cases with slightly denser LLSVPs are similar, although the geoid amplitudes are lower for the latter. However, with more dense LLSVPs, the slabs cannot perturb them and no plumes are generated. Plumes arise as thermal instabilities from the edges of the LLSVPs, when cold and viscous slabs perturb them. While our predicted plume locations are consistent with the observed hotspot locations, matching the plume structures in tomography models is difficult. These plumes are essential in fitting the finer features of the observed geoid. In longer-duration models, more voluminous subducted material reaches the CMB, which tends to erode the LLSVPs significantly, and yields a poor fit to the observed geoid. Our results suggest that with the presence of a thin, moderately weak layer below 660 km, a slightly dense LLSVP, and Clapeyron slope of -2.5 MPa/K, the velocity anomalies in seismic tomography and the long-wavelength geoid can be matched well. One of the limitations of our models is that the assimilated plate motion history may be too short to overcome arbitrary initial conditions effects. Also, assimilated true plate velocities in our models may not represent the true convective vigor of the Earth.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"90 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141508208","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yu Hong, Jianghai Xia, Hao Zhang, Jingyin Pang, Bo Guan, Jie Zhou, Yulong Ma
Summary Passive surface-wave methods have found extensive application in near-surface investigation due to their benefits of low costs, noninvasiveness, and high accuracy. Linear arrays are usually adopted in urban environments for their convenience and efficiency. However, the distribution of noise sources in densely populated urban areas varies rapidly in time and space, making it challenging to estimate accurate dispersion spectra using a linear array. To solve this problem, we propose a polarization analysis-based azimuthal correction method. We first obtain the azimuth of each segment by calculating the correlation coefficient of three-component ambient noise data. The normalized correlation coefficient is then applied for quality control to select reliable segments. For selected segments, the overestimated velocity caused by directional sources are corrected to obtain accurate dispersion spectra. A synthetic test is conducted to demonstrate the feasibility of our method. Compared with the dispersion spectra obtained without any correction, the dispersion spectra obtained following the suggested scheme are more consistent with the theoretical dispersion curves. Two real-world examples at crossroads show the superiority of the proposed technique in obtaining higher-resolution dispersion energy and more accurate phase velocities. In addition, our approach can attenuate the artifacts and improve the dispersion measurements.
{"title":"Azimuth correction for passive surface wave dispersion based on polarization analysis","authors":"Yu Hong, Jianghai Xia, Hao Zhang, Jingyin Pang, Bo Guan, Jie Zhou, Yulong Ma","doi":"10.1093/gji/ggae232","DOIUrl":"https://doi.org/10.1093/gji/ggae232","url":null,"abstract":"Summary Passive surface-wave methods have found extensive application in near-surface investigation due to their benefits of low costs, noninvasiveness, and high accuracy. Linear arrays are usually adopted in urban environments for their convenience and efficiency. However, the distribution of noise sources in densely populated urban areas varies rapidly in time and space, making it challenging to estimate accurate dispersion spectra using a linear array. To solve this problem, we propose a polarization analysis-based azimuthal correction method. We first obtain the azimuth of each segment by calculating the correlation coefficient of three-component ambient noise data. The normalized correlation coefficient is then applied for quality control to select reliable segments. For selected segments, the overestimated velocity caused by directional sources are corrected to obtain accurate dispersion spectra. A synthetic test is conducted to demonstrate the feasibility of our method. Compared with the dispersion spectra obtained without any correction, the dispersion spectra obtained following the suggested scheme are more consistent with the theoretical dispersion curves. Two real-world examples at crossroads show the superiority of the proposed technique in obtaining higher-resolution dispersion energy and more accurate phase velocities. In addition, our approach can attenuate the artifacts and improve the dispersion measurements.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"147 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141508207","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary We present a novel strategy for performing joint inversion with guided fuzzy c-means (GFCM) clustering coupling and apply it to electrical resistivity tomography (ERT) and ambient noise surface wave (ANSW) data. To accurately extract a priori clustering information, we use density peak clustering (DPC) rather than fuzzy c-means (FCM). The number and centres of resistivity and shear-wave velocity a priori clusters are extracted by DPC and then used to guide the joint inversion with the GFCM clustering coupling of ERT and ANSW data. Synthetic and field data are used to evaluate the flow and algorithm of DPC-GFCM clustering joint inversion. The results of synthetic examples show that the models recovered by the DPC-GFCM clustering joint inversion are nearly the same as the true models and are more accurate than those inverted using individual inversion and FCM-GFCM clustering joint inversion. In the field case, the depths of the stratigraphic interfaces shown in the resistivity and shear-wave velocity models inverted by DPC-GFCM clustering joint inversion are nearly consistent with those from the drilling data. In contrast, the strata recovered by the individual inversion and FCM-GFCM clustering joint inversion significantly differ from the drilling results. Both the synthetic and field examples verify the effectiveness of the DPC-GFCM clustering coupling method used for the joint inversion of ERT and ANSW data acquired from the near surface with strong heterogeneity. This novel approach can also be applied to other types of geophysical data.
{"title":"Joint inversion of ERT and ambient noise surface wave data with DPC-guided fuzzy c-means clustering for near-surface imaging","authors":"Zhanjie Shi, Chao Wang","doi":"10.1093/gji/ggae227","DOIUrl":"https://doi.org/10.1093/gji/ggae227","url":null,"abstract":"Summary We present a novel strategy for performing joint inversion with guided fuzzy c-means (GFCM) clustering coupling and apply it to electrical resistivity tomography (ERT) and ambient noise surface wave (ANSW) data. To accurately extract a priori clustering information, we use density peak clustering (DPC) rather than fuzzy c-means (FCM). The number and centres of resistivity and shear-wave velocity a priori clusters are extracted by DPC and then used to guide the joint inversion with the GFCM clustering coupling of ERT and ANSW data. Synthetic and field data are used to evaluate the flow and algorithm of DPC-GFCM clustering joint inversion. The results of synthetic examples show that the models recovered by the DPC-GFCM clustering joint inversion are nearly the same as the true models and are more accurate than those inverted using individual inversion and FCM-GFCM clustering joint inversion. In the field case, the depths of the stratigraphic interfaces shown in the resistivity and shear-wave velocity models inverted by DPC-GFCM clustering joint inversion are nearly consistent with those from the drilling data. In contrast, the strata recovered by the individual inversion and FCM-GFCM clustering joint inversion significantly differ from the drilling results. Both the synthetic and field examples verify the effectiveness of the DPC-GFCM clustering coupling method used for the joint inversion of ERT and ANSW data acquired from the near surface with strong heterogeneity. This novel approach can also be applied to other types of geophysical data.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"16 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141514192","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary On 6 February 2023, an Mw 7.8 left-lateral strike-slip fault earthquake occurred on the East Anatolian Fault Zone (EAFZ) in Türkiye. This study examined the spatial variation of the stress field around Türkiye better to understand the generation process of this event. We first combined focal mechanisms around Türkiye, created a dataset consisting of 2984 focal mechanisms, and conducted stress tensor inversion. The results showed that the maximum compressional axis near the EAFZ was oriented north-south and slightly varied along the strike. Moreover, the relative magnitude of north-south compressional stress gradually increases from south to north, and the stress regime changes from a normal fault stress regime to a strike-slip fault regime. The stress change caused by the Mw 7.8 mainshock does not explain this lateral pattern, implying that this stress regime transition existed before the mainshock. This suggests that shear stress on the EAFZ was low in this southern segment because it was unfavourably oriented to the regional stress field. Previous studies have reported that the Mw 7.8 mainshock rupture started at a splay fault, first propagated through the central and northern segments and then backpropagated with a time delay toward the southern segment, where it caused a significant but relatively small slip. The preexisting along-strike shear stress variation on the fault may have contributed to the smaller and delayed coseismic slip in the southern segment than in the central and northern segments. Moreover, the mainshock rupture possibly caused stress rotation locally near the central segment where the magnitudes of the vertical and north-south compressional stresses were almost equal.
{"title":"Spatial Variation in Stress Orientation around Türkiye: Rupture Propagation across the Stress Regime Transition in the 2023 Mw 7.8 Türkiye Earthquake","authors":"Keisuke Yoshida","doi":"10.1093/gji/ggae230","DOIUrl":"https://doi.org/10.1093/gji/ggae230","url":null,"abstract":"Summary On 6 February 2023, an Mw 7.8 left-lateral strike-slip fault earthquake occurred on the East Anatolian Fault Zone (EAFZ) in Türkiye. This study examined the spatial variation of the stress field around Türkiye better to understand the generation process of this event. We first combined focal mechanisms around Türkiye, created a dataset consisting of 2984 focal mechanisms, and conducted stress tensor inversion. The results showed that the maximum compressional axis near the EAFZ was oriented north-south and slightly varied along the strike. Moreover, the relative magnitude of north-south compressional stress gradually increases from south to north, and the stress regime changes from a normal fault stress regime to a strike-slip fault regime. The stress change caused by the Mw 7.8 mainshock does not explain this lateral pattern, implying that this stress regime transition existed before the mainshock. This suggests that shear stress on the EAFZ was low in this southern segment because it was unfavourably oriented to the regional stress field. Previous studies have reported that the Mw 7.8 mainshock rupture started at a splay fault, first propagated through the central and northern segments and then backpropagated with a time delay toward the southern segment, where it caused a significant but relatively small slip. The preexisting along-strike shear stress variation on the fault may have contributed to the smaller and delayed coseismic slip in the southern segment than in the central and northern segments. Moreover, the mainshock rupture possibly caused stress rotation locally near the central segment where the magnitudes of the vertical and north-south compressional stresses were almost equal.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"121 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141514190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Chunhong Wu, Xinwen Su, Chuang Xu, Guangyu Jian, Jinbo Li
Summary During the inversion of seafloor topography (ST) using the backpropagation neural network (BPNN), the random selection of parameters may decrease the accuracy. To address this issue and achieve a more efficient global search, this paper introduces a genetic algorithm-backpropagation (GA-BP) neural network. Benefiting from the global search and parallel computing capabilities of the GA, this study refines the seafloor topography of the South China Sea using multi-source gravity data. The results indicate that the GA-BP model, with a root mean square (RMS) value of 126.0 m concerning ship-measured water depths. It is noteworthy that when dealing with regions characterized by sparse survey line distributions, the GA-BP neural network stronger robustness compared to BPNN, showing less sensitivity to the distribution of survey data. Furthermore, the paper explores the influence of different data preprocessing methods on the neural network inversion of sea depths. This research introduces an optimization algorithm that reduces instability during BPNN initialization, resulting in a more accurate prediction of seafloor topography.
{"title":"Seafloor topography refinement from multi-source data using genetic algorithm - backpropagation neural network","authors":"Chunhong Wu, Xinwen Su, Chuang Xu, Guangyu Jian, Jinbo Li","doi":"10.1093/gji/ggae229","DOIUrl":"https://doi.org/10.1093/gji/ggae229","url":null,"abstract":"Summary During the inversion of seafloor topography (ST) using the backpropagation neural network (BPNN), the random selection of parameters may decrease the accuracy. To address this issue and achieve a more efficient global search, this paper introduces a genetic algorithm-backpropagation (GA-BP) neural network. Benefiting from the global search and parallel computing capabilities of the GA, this study refines the seafloor topography of the South China Sea using multi-source gravity data. The results indicate that the GA-BP model, with a root mean square (RMS) value of 126.0 m concerning ship-measured water depths. It is noteworthy that when dealing with regions characterized by sparse survey line distributions, the GA-BP neural network stronger robustness compared to BPNN, showing less sensitivity to the distribution of survey data. Furthermore, the paper explores the influence of different data preprocessing methods on the neural network inversion of sea depths. This research introduces an optimization algorithm that reduces instability during BPNN initialization, resulting in a more accurate prediction of seafloor topography.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"188 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141514191","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary Mode conversion of P waves at the boundary between Earth's crust and upper mantle, when analyzed using receiver functions (RFs), allows characterization of Earth structure where seismic station density is high and earthquake sources are favorably distributed. We applied two ensemble decision tree algorithms – Random Forest (RanFor) and eXtreme Gradient Boost (XGBoost) – to synthetic and real RF data to assess these machine learning techniques' potential for crustal imaging when available data are sparse. The synthetic RFs, entailing both sharp increases in seismic velocity across the Moho and gradational Moho structures, calculated with and without added random noise, correspond to idealized crustal structures: a dipping Moho, Moho offset by crustal-scale faults, anti- and synform Moho structures and combinations of these. The RanFor/XGBoost algorithm recovers input structures well regardless of event-station distributions. Useful crustal and upper mantle seismic velocities can also be determined using RanFor and XGBoost, making it possible to image crustal thickness and P and S wave velocities simultaneously from receiver functions alone. We applied the trained RanFor/XGBoost to receiver functions determined from real seismic data recorded in the contiguous U.S., producing a map of the Moho and P and S wave velocities of the lowermost crust and uppermost mantle. Use of XGBoost, which evaluates residuals between input RFs and ground-truth to update the decision tree using the gradient of a penalty function, improves the crustal thickness estimates.
摘要 利用接收函数（RF）分析地壳和上地幔边界处 P 波的模式转换，可以确定地震台站密度高且震源分布合理的地球结构特征。我们将随机森林（RanFor）和极端梯度提升（XGBoost）这两种决策树算法应用于合成和真实射频数据，以评估这些机器学习技术在可用数据稀少的情况下用于地壳成像的潜力。合成射频数据包括莫霍区地震速度的急剧增加和渐变莫霍区结构，计算时添加和不添加随机噪声，这些数据对应于理想化的地壳结构：倾斜莫霍区、被地壳尺度断层抵消的莫霍区、反形和合形莫霍区结构以及这些结构的组合。无论事件站分布如何，RanFor/XGBoost 算法都能很好地恢复输入结构。使用 RanFor 和 XGBoost 算法还能确定有用的地壳和上地幔地震速度，从而有可能仅通过接收函数就能同时对地壳厚度以及 P 波和 S 波速度进行成像。我们将训练有素的 RanFor/XGBoost 应用于从美国毗连地区记录的真实地震数据中确定的接收函数，绘制出了莫霍线图以及最下部地壳和最上部地幔的 P 波和 S 波速度。XGBoost 可以评估输入射频和地面实况之间的残差，利用惩罚函数的梯度更新决策树，从而改进地壳厚度估算。
{"title":"Use of Decision Tree Ensembles for Crustal Structure Imaging from Receiver Functions","authors":"Yitan Wang, R M Russo, Yuanhang Lin","doi":"10.1093/gji/ggae226","DOIUrl":"https://doi.org/10.1093/gji/ggae226","url":null,"abstract":"Summary Mode conversion of P waves at the boundary between Earth's crust and upper mantle, when analyzed using receiver functions (RFs), allows characterization of Earth structure where seismic station density is high and earthquake sources are favorably distributed. We applied two ensemble decision tree algorithms – Random Forest (RanFor) and eXtreme Gradient Boost (XGBoost) – to synthetic and real RF data to assess these machine learning techniques' potential for crustal imaging when available data are sparse. The synthetic RFs, entailing both sharp increases in seismic velocity across the Moho and gradational Moho structures, calculated with and without added random noise, correspond to idealized crustal structures: a dipping Moho, Moho offset by crustal-scale faults, anti- and synform Moho structures and combinations of these. The RanFor/XGBoost algorithm recovers input structures well regardless of event-station distributions. Useful crustal and upper mantle seismic velocities can also be determined using RanFor and XGBoost, making it possible to image crustal thickness and P and S wave velocities simultaneously from receiver functions alone. We applied the trained RanFor/XGBoost to receiver functions determined from real seismic data recorded in the contiguous U.S., producing a map of the Moho and P and S wave velocities of the lowermost crust and uppermost mantle. Use of XGBoost, which evaluates residuals between input RFs and ground-truth to update the decision tree using the gradient of a penalty function, improves the crustal thickness estimates.","PeriodicalId":12519,"journal":{"name":"Geophysical Journal International","volume":"83 1","pages":""},"PeriodicalIF":2.8,"publicationDate":"2024-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141514189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：