Table of contents

In this paper, the response of different synthetic models to geoelectrical multi-electrode surveys is studied. The models considered are related to two main problems, which are very common in geophysical research regarding hydrogeology and engineering. The first class of models represents buried walls, similar archaeological remains or remains of buried foundations; the other class corresponds to a sea-water intrusion of a fresh water aquifer, which is generally studied in hydrogeophysics. A set of 2D simulations, starting from the synthetic models, was carried out to compare the behaviour of the different arrays when acquiring measurements of electrical resistivity tomography. For each model, the apparent resistivity data—and relative pseudo-sections—were calculated for common electrode arrays (Wenner, dipole–dipole, Wenner–Schlumberger). Furthermore, a 'non-classical' configuration (i.e. the linear grid) was also tested. The synthetic data, after adding different levels of Gaussian noise, were inverted using RES2DINV software; then the interpretative models were compared with the initial synthetic models using different parameters to estimate the quality of the matching. Finally, the comparison between the results obtained using the various arrays is presented. Furthermore, the effectiveness of the various arrays is evaluated for each problem, also taking into account some other characteristics of the arrays, including the associated practical advantages in time consumption and noise level. Results on synthetic data were also confirmed by two field tests: one at an archaeological site survey and one at a coastal aquifer study.

This is a case study on the application of inverse-Q filtering to improve the resolution of 3D seismic data, for the characterization of tight-sand gas reservoirs. When seismic waves propagate through multiple tight-sand layers in the subsurface media, the energy of high-frequency components is absorbed, and the wavelet shape is distorted. Stabilized inverse-Q filtering is able to simultaneously compensate the amplitude and correct the phase of seismic waveforms. After application to 3D seismic data, the frequency bandwidth has been increased by about 10 Hz, and the width of the wavelet has been narrowed, so that we are able to identify reflections of thin sand layers clearly. Due to the phase correction in inverse-Q filtering, filtered seismic traces can match the synthetic traces at well locations. Because of the high signal-to-noise ratio with the stabilization scheme, low-amplitude zones of interest corresponding to target high-fracture areas can be easily identified, and the detail within the anomalies can also be observed. Finally, spatial variations of tight-sand layers are depicted in the inversion profile with high resolution.

The modular neural network (MNN) inversion method has been used for inversion of self-potential (SP) data anomalies caused by 2D inclined sheets of infinite horizontal extent. The analysed parameters are the depth (h), the half-width (a), the inclination (α), the zero distance from the origin (x_o) and the polarization amplitude (k). The MNN inversion has been first tested on a synthetic example and then applied to two field examples from the Surda area of Rakha mines, India, and Kalava fault zone, India. The effect of random noise has been studied, and the technique showed satisfactory results. The inversion results show good agreement with the measured field data compared with other inversion techniques in use.

Velocity analysis is always a critical step during seismic processing due to its influence on the final quality of the data. Over the last few decades, residual migration velocity analysis (RMVA) as a popular tool to estimate a velocity–depth model has been an area of active research. Previous works on the RMVA have been computationally intensive to be applied in 3D cases. In this paper, we propose a 3D synthesized areal shot-record RMVA method with controlled illumination. In the scheme, the interval velocities are updated using the residual analysis in the plane wave domain only with the layer-stripping residual correction rather than the layer-stripping prestack depth migration (PSDM). The synthesized areal shot-record migration is performed in the plane wave domain. This is consistent with the condition of the velocity correction formulae. The common image gather (CIG) sections can be extracted directly from the migrated data. To avoid the propagation distortion of the plane wave source wavefield through an inhomogeneous subsurface, the controlled illumination technique is employed. Results on both synthetic data and field data demonstrate that the method is practical and efficient.

We have developed a semi-automatic method to determine the depth and shape (shape factor) of a buried structure from second moving average residual self-potential anomalies obtained from observed data using filters of successive window lengths. The method involves using a relationship between the depth and the shape to source and a combination of windowed observations. The relationship represents a parametric family of curves (window curves). For a fixed window length, the depth is determined for each shape factor. The computed depths are plotted against the shape factors, representing a continuous monotonically increasing curve. The solution for the shape and depth is read at the common intersection of the window curves. The validity of the method is tested on a synthetic example with and without random errors and on two field examples from Turkey and Germany. In all cases examined, the depth and the shape solutions obtained are in very good agreement with the true ones.

An ellipse evolving common reflection point (CRP) is an innovative stack method. Under some velocity distribution, we can obtain a real zero-offset section by projecting seismic signals to isochrones along an elliptic trajectory. This paper introduces this method and its velocity analysis. By theoretical model and real seismic data processing, the ellipse evolving CRP velocity analysis more accurately estimates the CRP stack velocity of a complex geology section. The method also more clearly detects an interval velocity anomaly. It is a valuable method of reservoir prediction and oil and gas detection.

In this paper, the correlation between shear wave velocity and standard penetration test blow counts (SPT-N) is investigated. The study focused primarily on the correlation of SPT-N and shear wave velocity (V_s) for several soil categories: all soils, sand, silt and clay-type soils. New empirical formulae are suggested to correlate SPT-N and V_s, based on a dataset collected in a part of Eskişehir settlement in the western central Anatolia region of Turkey. The formulae are based on geotechnical soundings and active and passive seismic experiments. The new and previously suggested formulae showing correlations between uncorrected SPT-N and V_s have been compared and evaluated by using the same dataset. The results suggest that better correlations in estimation of V_s are acquired when the uncorrected blow counts are used. The blow count is a major parameter and the soil type has no significant influence on the results. In cohesive soils, the plasticity contents and, in non-cohesive soils except for gravels, the graded contents have no significant effect on the estimation of V_s. The results support most of the conclusions of earlier studies. These practical relationships developed between SPT-N and V_s should be used with caution in geotechnical engineering and should be checked against measured V_s.

The Yongjang mine is an Au–Ag deposit near Masan, located at the southernmost tip of the Korean Peninsula. The deposit lies within Cretaceous sedimentary rocks and contains many quartz veins which contain elements such as gold and silver, and sulfides. In the mine, the Yongjang, En and Ansan quartz veins have been found to be gold bearing. These veins have thicknesses of 2–40 cm and extents of 100–260 m. Electrical resistivity surveys were conducted to clarify the location of gold deposits at both prospect and detailed scales. Apparent resistivity data were collected with a dipole–dipole array on the ground surface and in boreholes, and with a pole–dipole array for surface-to-borehole surveys. The datasets derived from three-dimensional inversion of apparent resistivities are quite effective at delineating the geological structures related to gold-bearing quartz veins. These appear as a low-resistivity anomaly because almost all of the gold mineralization occurs in fractured areas associated with faults or shear zones. The surface-to-borehole survey had better resolution than the surface dipole–dipole survey when imaging gold-bearing quartz veins. The low-resistivity anomalies indicating the Yongjang and Ansan veins extend nearly vertically to sea level and dip steeply below sea level. They run NW–SE parallel to each other at a distance of about 70 m. The En vein is imaged near the Yonjang vein with a strike direction of N60°–70° W and a dip angle of about 45°.

Based on the requirement of probability tomography, we developed two-directional pole–pole measurement on a 2D electrical imaging survey which is able to extract all second electrical field profiles to participate in probability tomography. The concept of charge occurrence probability (COP) is defined as the average of cross-correlation of the space domain scanning (SDS) function with the second electrical field being generated by a whole set of point current sources. To quantitatively analyse the results of the tomography, the COP values are normalized. Finally, in order to test the effectiveness of the tomography, we use a finite element algorithm to synthesize the potential data for 2D models, and the profiles of the second electrical field are extracted by eliminating the primary field from the total field. Using the second electrical field of all profiles, the geometric patterns of normalized COP (NCOP) values are reconstructed by means of multiple superimposed probability tomography. The results are quite satisfactory.

A blind deconvolution technique using a modified higher order statistics (HOS)-based eigenvector algorithm (EVA) is presented in this paper. The main purpose of the technique is to enable the processing of low SNR short length seismograms. In our study, the seismogram is assumed to be the output of a mixed phase source wavelet (system) driven by a non-Gaussian input signal (due to earth) with additive Gaussian noise. Techniques based on second-order statistics are shown to fail when processing non-minimum phase seismic signals because they only rely on the autocorrelation function of the observed signal. In contrast, existing HOS-based blind deconvolution techniques are suitable in the processing of a non-minimum (mixed) phase system; however, most of them are unable to converge and show poor performance whenever noise dominates the actual signal, especially in the cases where the observed data are limited (few samples). The developed blind equalization technique is primarily based on the EVA for blind equalization, initially to deal with mixed phase non-Gaussian seismic signals. In order to deal with the dominant noise issue and small number of available samples, certain modifications are incorporated into the EVA. For determining the deconvolution filter, one of the modifications is to use more than one higher order cumulant slice in the EVA. This overcomes the possibility of non-convergence due to a low signal-to-noise ratio (SNR) of the observed signal. The other modification conditions the cumulant slice by increasing the power of eigenvalues of the cumulant slice, related to actual signal, and rejects the eigenvalues below the threshold representing the noise. This modification reduces the effect of the availability of a small number of samples and strong additive noise on the cumulant slices. These modifications are found to improve the overall deconvolution performance, with approximately a five-fold reduction in a mean square error (MSE) and a six-fold reduction in convolution noise of 10 dB of the SNR.