- Home
- A-Z Publications
- Near Surface Geophysics
- Previous Issues
- Volume 19, Issue 4, 2021
Near Surface Geophysics - Volume 19, Issue 4, 2021
Volume 19, Issue 4, 2021
-
-
A hybrid residual neural network–Monte Carlo approach to invert surface wave dispersion data
Authors Mattia Aleardi and Eusebio StucchiABSTRACTSurface‐wave inversion is a non‐linear and ill‐conditioned problem usually solved through deterministic or global optimization approaches. Here, we present an alternative method based on machine learning. Under the assumption of a local one‐dimensional model, we train a residual neural network to predict the non‐linear mapping between the full dispersion image and the model space, parameterized in terms of shear wave velocity and layer thicknesses. On the one hand, compared to standard convolutional neural networks, the residual network prevents the vanishing gradient problem when training a deep network. On the other hand, the use of the full dispersion image avoids the time‐consuming and often ambiguous picking procedure and allows considering higher modes in the inversion framework. One key aspect of any machine learning inversion strategy is the definition of an appropriate training set. In this case, the models forming the training and validation examples are uniformly drawn from previously defined ranges that cover a wide range of possible near‐surface layered Vs models. The reflectivity method constitutes the forward modelling operator that converts the model parameters into the observed shot gathers. The inversion also includes a Monte Carlo simulation strategy that propagates onto the model space the uncertainties related to noise in the data and the modelling error introduced by the network approximation. We first discuss synthetic inversions to assess the applicability of the proposed method and to analyse the effect of erroneous model parameterizations. The inversion results are also benchmarked with those provided by a more standard approach in which the particle swarm optimization algorithm inverts the fundamental mode only. Then, we discuss a field data application. Our tests confirm that the residual neural network inversion provides accurate model estimations and reliable uncertainty appraisals. One of the main benefits of the proposed approach is that once the network is trained it provides the near‐surface shear wave velocity profile in near real‐time.
-
-
-
Memory‐efficient frequency‐domain Gauss–Newton method for wave‐equation first‐arrival traveltime inversion
Authors Jianhua Wang and Liangguo DongABSTRACTWave‐equation traveltime inversion (WTI) can be used to automatically obtain a background near‐surface velocity model (NSM), which overcomes the high‐frequency approximation in ray theory. It is generally implemented in the time domain. However, the commonly used gradient‐based optimisation methods (such as the steepest‐descent method) in WTI have a low convergence rate and may yield less accurate results within limited iterations in geologically complex regions. To increase the convergence rate and improve the inversion accuracy, we propose a frequency‐domain truncated Gauss–Newton first‐arrival wave‐equation traveltime inversion (GN‐WTI) method to retrieve the background NSM. As only a few frequencies are used for inversion, the proposed frequency‐domain WTI method significantly reduces the computational memory requirements by more than two orders of magnitude in comparison with the conventional time‐domain WTI method. Therefore, the proposed method is especially advantageous for the building of large three‐dimensional models. In this GN‐WTI method, according to the derived explicit traveltime residual kernel, the gradient and Hessian vector products can be computed efficiently using an elegant and improved scattering integral approach as long as the source‐side wavefields and non‐redundant receiver‐side Green's functions are computed and stored in advance. The conjugate gradient approach is used to solve the Gauss–Newton normal equation to obtain the Gauss–Newton direction in inner loops. Here, the Gauss–Newton Hessians of the ray‐based traveltime inversion and WTI are compared to demonstrate the advantages of WTI. The trial runs with a simple periodic velocity model example showed that the proposed GN‐WTI method outperforms the WTI method when using the steepest‐descent and limited‐memory Broyden–Fletcher–Goldfarb–Shanno approaches in terms of the convergence rate and inversion accuracy. A complex Marmousi model was further used to illustrate the effectiveness of GN‐WTI. The proposed method should be beneficial in near‐surface velocity model building.
-
-
-
Microseismic monitoring of rockbursts with the ensemble Kalman filter
Authors A. Cecilia Dip, Bernard Giroux and Erwan GloaguenABSTRACTWe introduce an algorithm to monitor seismic velocity changes associated with rockbursts in mines, through microseismic monitoring. Rockbursts are extreme events resulting from the complex interaction between mining activities and geology, and represent a significant threat to mines. In recent years, the use of passive seismic monitoring for mine safety and productivity has progressed substantially, aiming to understand and predict this hazard. In this work, additional value is given to microseismic monitoring, using it to map temporal changes of seismic velocity in the rock mass that can potentially be associated with stress changes leading to rockbursts. An application of the ensemble Kalman filter is presented for assimilating travel times of seismic P and S waves in a fast and efficient way in order to update the mine's velocity model. Combining sequential Gaussian simulation and ensemble Kalman filter techniques, we were able to monitor the occurrence of velocity changes underground. The proposed approach allows zones to be highlighted where the rock mass is under stress and where potential risk can be expected. The performance of the method was first tested on several synthetic scenarios and, subsequently, on a real 3D case of a deep mine in Canada. The application on the real data set allowed mapping a change in velocity in the area where a rockburst occurred four hours afterwards.
-
-
-
Applicability of a proposed groundwater level determination approach for the K‐NET in Japan
More LessABSTRACTThe evaluation of groundwater presence by means of seismic wave velocities is of great concern for many hydrological related engineering applications. This work introduces a simple basic procedure to accurately and systematically identify groundwater levels using P‐wave and S‐wave velocities. This procedure is defined as a seismic reflectivity parameter and based on the seismic reflection coefficients of P‐wave and S‐wave at the unsaturated/saturated interfaces (i.e., groundwater level). The present procedure is applied to 1045 K‐NET sites in Japan. Observed velocity structures, accompanied by their bulk densities and detailed lithology, for the upper 20 m depth are obtained from the National Research Institute for Earth Science and Disaster Prevention, Japan. Calculations are carried out considering three distinct lithological conditions – gravel, sand and clay–silt sediments – in order to evaluate the ability of this approach (seismic reflectivity parameter) to detect the groundwater level and assess its practical limitations. The increase of the seismic reflectivity parameter value varies substantially with Poisson's ratio, resembling the relationship between and Poisson's ratio in gravel, sand and clay–silt sediments. However, positive seismic reflectivity parameter values indicate groundwater presence, whereas negative seismic reflectivity parameter values indicate lithological variation. The seismic reflectivity parameter can also be used as a relative indicator for groundwater saturation levels. Fair comparisons are established between the seismic reflectivity parameter and the water seismic index technique. However, the seismic reflectivity parameter approach shows consistency with previous theoretical and experimental studies related to variations of and in saturated conditions, whereas the water seismic index shows inconsistency with these previous theoretical and experimental studies and an inability in determining saturation levels and detecting groundwater levels.
-
-
-
Improving crosshole ground‐penetrating radar full‐waveform inversion results by using progressively expanded bandwidths of the data
Authors Zhen Zhou, Anja Klotzsche and Harry VereeckenABSTRACTIn the last decade, time‐domain crosshole ground‐penetrating radar full‐waveform inversion has been applied to several different test sites and has improved the resolution and reconstruction of subsurface properties. The full‐waveform inversion requires several diligent executed pre‐processing steps to guarantee a successful inversion and to minimize the risk of being trapped in a local minimum. Thereby, one important aspect is the starting models of the full‐waveform inversion. Generally, adequate starting models need to fulfil the half‐wavelength criterion, which means that the modelled data based on the starting models need to be within half of the wavelength of the measured data in the entire investigation area. Ray‐based approaches can provide such starting models, but in the presence of high contrast layers, such results do not always fulfil this criterion and need to be improved and updated. Therefore, precise and detailed data processing and a good understanding of experimental ground‐penetrating radar data are necessary to avoid erroneous full‐waveform inversion results. Here, we introduce a new approach, which improves the starting model problem and is able to enhance the reconstruction of the subsurface medium properties. The new approach tames the non‐linearity issue caused by high contrast complex media, by applying bandpass filters with different passband ranges during the inversion to the modelled and measured ground‐penetrating radar data. Thereby, these bandpass filters are considered for a certain number of iterations and are progressively expanded to the selected maximum frequency bandwidth. The resulting permittivity full‐waveform inversion model is applied to update the effective source wavelet and is used as an updated starting model in the full‐waveform inversion with the full bandwidth data. This full‐waveform inversion is able to enhance the reconstruction of the permittivity and electrical conductivity results in contrast to the standard full‐waveform inversion results. The new approach has been applied and tested on two synthetic case studies and an experimental data set. The field data were additionally compared with cone penetration test data for validation.
-
-
-
Electrical resistivity tomographic detection of the hidden Thakek fault, Northeast Thailand
ABSTRACTThe Khorat plateau covers most of northeast Thailand along with parts of southern Laos. Recent earthquake hazard assessment and other studies of the regional Thakek fault have rarely appeared in publications and reports. The Thakek fault zone in the Laos region includes three main principal fault planes, but it rapidly fades out when the fault passes through Thailand. A geophysical hidden fault study in the Khorat plateau is herein performed in Bung Kan and Nakhon Phanom provinces, Thailand, where the surface expression is difficult to document through conventional geological surveys. Some fault traces have been identified by satellite image interpretation and digital elevation model analysis. Geological field investigation and electrical resistivity tomography surveys were carried out in order to determine an optimal site for trenching. The electrical resistivity tomography results show zones of high and low resistivity interpreted as coarse sediment layers and mudstone, respectively. The maximum vertical fault offset occurring along the fault trace interpreted from the three‐dimensional electrical resistivity tomographic slices is 40 ± 5 m. Based on the electrical resistivity tomography results, a trench was excavated to directly investigate patterns of faulting and deformation. Based on structural geological observations at the trench wall, indications of faulting have been tentatively identified. In particular, faulting appears to have displaced mudstone over Quaternary sediments with vertical offsets >3.5 m cutting the youngest unit. The result has implications for seismic hazard assessment in northeast Thailand.
-
Volumes & issues
-
Volume 22 (2024)
-
Volume 21 (2023)
-
Volume 20 (2022)
-
Volume 19 (2021)
-
Volume 18 (2020)
-
Volume 17 (2019)
-
Volume 16 (2018)
-
Volume 15 (2017)
-
Volume 14 (2015 - 2016)
-
Volume 13 (2015)
-
Volume 12 (2013 - 2014)
-
Volume 11 (2013)
-
Volume 10 (2012)
-
Volume 9 (2011)
-
Volume 8 (2010)
-
Volume 7 (2009)
-
Volume 6 (2008)
-
Volume 5 (2007)
-
Volume 4 (2006)
-
Volume 3 (2005)
-
Volume 2 (2004)
-
Volume 1 (2003)