Vertical Seismic Profiling Measurements Research Articles

A Novel Method for Determining Geophone Orientations From Zero-Offset VSP Data Constrained by Scalar Field

In vertical seismic profile (VSP) acquisition, the orientation of the three-component geophones in a borehole is often not aligned due to random rotation of the wireline receiver array. Knowing the geophone orientation is crucial for rotating the waveforms back to the position where the geophones are assumed to be aligned. This procedure is critical in three-component VSP data processing as it provides the correct wavefields for subsequent VSP data imaging and inversion. For zero-offset or near zero-offset VSP measurements, the direct P-wave propagates nearly vertically, and the horizontal geophones receive only a small fraction of the direct P-wave energy. The first-arrival in the horizontal component data is often of low signal-to-noise ratio (SNR). As a result, it is difficult to apply conventional first-arrival–based methods for obtaining accurate geophone orientations. On the other hand, a seismic event comprising individual waveform would achieve maximum correlation if all corresponding geophones are aligned. The geometric characteristics, that is, the slope and continuity of a seismic event in the vector wavefield are the same with those of the modules of the vector wavefield. The latter, also known as the scalar wavefield, can be used for scanning geophone orientations based on waveform correlation. In this study, we propose to use the scalar field to extract the slope of seismic events. The orientation of the individual geophone was calculated with the constraint of the slope. We use both synthetic and field data to demonstrate the effectiveness and applicability of the proposed method. The results show that the new method can provide wavefields of horizontal component VSP data with much higher accuracy and resolution.

Deep Learning-Based P- and S-Wave Separation for Multicomponent Vertical Seismic Profiling

Vertical seismic profiling (VSP) helps to derive high-resolution images around the instrumented borehole and is a cost-effective technique for CO₂ storage monitoring. In routine VSP data processing, P- and S-wave separation is a crucial step to extract independent single-mode waves for accurate imaging and interpretation. Conventional wave mode separation involves tedious, subjective, and non-reproducible manual interventions, especially when dealing with complex geology. To better automate the process, we propose a data-driven deep learning-based P- and S-wave separation method. Our method adapts a fully convolutional neural network that simultaneously extracts P- and S-potential data from multicomponent VSP measurements. To reduce the enormous computational cost in wave simulation while constructing training datasets with sufficient kinematic and dynamic variations, we introduce virtual wellbores where synthetic VSP data sampling wide variations in seismic kinematics and dynamics are recorded using only a dozen elastic wave simulations on a single velocity model. We qualify the separation results both directly in data space and in image space after reverse time migration (RTM). Generalization tests on various synthetic models and their corresponding RTM images demonstrate that the proposed strategy provides sufficient sampling of the high-dimensional data space and essentially ensures successful applications of the trained neural network to similar yet different geological scenarios.

Orthotropic anisotropy analysis and parameter estimation from 3-D vertical seismic profile data

SUMMARY Orthotropic velocity models are commonly considered as to have the lowest complexity (or highest symmetry) that can realistically describe the combined effect of Earth’s velocity heterogeneity in the vertical and in the horizontal (radial) directions. Velocity heterogeneity in the vertical direction is commonly caused by horizontal stratification of subsurface rock layers with different elastic properties or caused intrinsically by microlayered rocks such as shales. In the horizontal direction, azimuthal variation of the Earth’s stress field can initially lead to preferential deformation of pore space in the rocks and eventual development of aligned (vertical) fractures that are commonly regarded as the main cause for seismic azimuthal anisotropy. Building an appropriate orthotropic velocity model that can explain seismic velocity variations in vertical planes and along any azimuthal direction can therefore enhance the quality of seismic data processing and imaging. This, however, requires knowledge of several anisotropy parameters defined in the three symmetry planes of the orthotropic model and also the orientation of these planes. A combination of P- and S-wave slowness and polarization data derived from vertical seismic profile (VSP) measurements has been routinely used to estimate parameters of less complex anisotropic media such as transverse isotropy with vertical axes of symmetry (VTI). The accuracy of anisotropy parameter estimation by these methods depends primarily on the availability of data that are used in the inversion and the size of anisotropy parameters. In this study, we develop a benchmark for the accuracy of orthotropic parameters that are estimated from 3-D VSP measurements using P-wave slowness only method. Through numerical analysis, we model a wide range of fracture-induced orthotropic symmetries by adding vertical fractures to a background VTI media and show that the accuracy of our orthotropic parameter estimation method is dependent on anisotropy in the host media, intensity of fracturing, noise level in slowness data and slowness data coverage in both the dip and azimuthal planes. We then expand our accuracy analysis work to multilayer media by examining the accuracy of P-wave slowness method for orthotropic anisotropy parameter estimation by inverting finite-difference data generated in a pseudo 3-D VSP experiment. We show that, for a typical 3-D VSP experiment, reasonable estimates of orthotropic parameters in the model can only be obtained with slowness data taken from Xmax/Z ≥ 1 and line azimuth interval ϕ ≤ 10° where Xmax is the maximum source to receiver offset and Z is the receiver depth. We finally describe how field VSP measurements can be inverted for orthotropic anisotropy parameters.

Underground Vertical Seismic Profiling with Conventional and Fiber-Optic Systems for Exploration in the Kylylahti Polymetallic Mine, Eastern Finland

Seismic reflection methods have been used for the exploration of mineral resources for several decades. However, despite their unmatched spatial resolution and depth penetration, they only have played a minor role in mineral discoveries so far. Instead, mining and exploration companies have traditionally focused more on the use of potential field, electric and electromagnetic methods. In this context, we present a case study of an underground Vertical Seismic Profiling (VSP) experiment, which was designed to image a (semi-)massive sulfide deposit located in the Kylylahti polymetallic mine in eastern Finland. For the measurement, we used a conventional VSP with three-component geophones and a novel fiber-optic Distributed Acoustic Sensing (DAS) system. Both systems were deployed in boreholes located nearby the target sulfide deposit, and used in combination with an active seismic source that was fired from within the underground tunnels. With this setup, we successfully recorded seismic reflections from the deposit and its nearby geological contrasts. The recording systems provided data with a good signal-to-noise ratio and high spatial resolution. In addition to the measurements, we generated a realistic synthetic dataset based on a detailed geological model derived from extensive drilling data and petrophysical laboratory analysis. Specific processing and imaging of the acquired and synthetic datasets yielded high-resolution reflectivity images. Joint analysis of these images and cross-validation with lithological logging data from 135 nearby boreholes led to successful interpretation of key geological contacts including the target sulfide mineralization. In conclusion, our experiment demonstrates the value of in-mine VSP measurements for detailed resource delineation in a complex geological setting. In particular, we emphasize the potential benefit of using fiber-optic DAS systems, which provide reflection data at sufficient quality with less logistical effort and a higher acquisition rate. This amounts to a lower total acquisition cost, which makes DAS a valuable tool for future mineral exploration activities.

Use of distributed fibre-optic sensing for marine seismic measurements

Heterodyne distributed vibration sensing (hDVS) and, more generically, distributed acoustic sensing (DAS) are technologies that rely on traditional fibre-optic cables to provide inline strain measurement. In recent years, hDVS has found application in seismic monitoring; in particular vertical seismic profiling (VSP) greatly benefits from the advantage of using permanent fibres cemented behind casing (Mateeva et al., 2013). VSP measurements were also demonstrated using fibre installed on production tubing (Barberan et al., 2012) and deployed within a wireline cable (Hartog et al., 2014). Another potential application for hDVS is flow monitoring, where, owing to the fibre’s large spatial coverage, it is possible to monitor vibrations and waves along the length of the well. Baihly et al. (2015) offered an insight on this use of hDVS. Furthermore, Molteni et al. (2016) evaluated microseismic monitoring as another compelling hDVS application. This paper evaluates the viability of using hDVS for distributed dynamic pressure sensing in marine seismic acquisition while reviewing the readiness of hDVS technology for this application. An initial concept of including hDVS in a streamer was disclosed by Kragh et al. (2012). Here, a possible design of a towed streamer using hDVS was discussed along with experimental results showing the potential of the technology to acquire pressure measurements in marine seismic surveys. We show the results of using hDVS to measure the pressure wavefield in two different experiments, and the results are compared with reference data acquired with traditional hydrophones. The use of a distributed measurement in towed streamers would provide a new type of measurement. The characterization of the pressure wavefield would be continuous along the cable with a denser spatial sampling. Similarly, straight fibres deployed along the streamer could characterize the tugging noise with increased spatial resolution. Furthermore, the use of cost-effective and lightweight fibres would greatly reduce the electronics embedded in the streamer and the total weight of the cable. The use of hDVS would also have new challenges – for example in how to segment the streamer into shorter sections without the fibre-optic connectors compromising the quality of the backscattered hDVS signal.

Imaging blended vertical seismic profiling data using full-wavefield migration in the common-receiver domain

For vertical-seismic-profiling (VSP) measurements, the use of blended acquisition, with time-overlapping shot records, can greatly reduce the downtime and, thereby, provide large cost savings. For directly imaging blended VSP measurements, we have used full-wavefield migration (FWM). FWM is an inversion-based imaging scheme that enables us to use any kind of complex source wavefield to estimate the subsurface reflectivity, using all the multiples (surface and internal) in the recorded data. The multiple scattering helps in improving the illumination as well as the vertical resolution of the image. In this scheme, active deblending is not required because the imaging process itself acts as a deblending procedure. We tested the potential of FWM to image blended VSP data, using simple and complex synthetic models. We clearly determined that using the primaries, surface multiples, and internal multiples enhanced the illumination away from the well trajectory, in which blending noise was suppressed due to the inversion scheme. We observed that some blending crosstalk noise leaked into the images with large blending factors. Such noise could be further reduced with additional constraints in the involved least-squares inversion process.

Finite-difference modelling to evaluate seismic P-wave and shear-wave field data

Abstract. High-resolution reflection seismic methods are an established non-destructive tool for engineering tasks. In the near surface, shear-wave reflection seismic measurements usually offer a higher spatial resolution in the same effective signal frequency spectrum than P-wave data, but data quality varies more strongly. To discuss the causes of these differences, we investigated a P-wave and a SH-wave seismic reflection profile measured at the same location on the island of Föhr, Germany and applied seismic reflection processing to the field data as well as finite-difference modelling of the seismic wave field. The simulations calculated were adapted to the acquisition field geometry, comprising 2 m receiver distance (1 m for SH wave) and 4 m shot distance along the 1.5 km long P-wave and 800 m long SH-wave profiles. A Ricker wavelet and the use of absorbing frames were first-order model parameters. The petrophysical parameters to populate the structural models down to 400 m depth were taken from borehole data, VSP (vertical seismic profile) measurements and cross-plot relations. The simulation of the P-wave wave-field was based on interpretation of the P-wave depth section that included a priori information from boreholes and airborne electromagnetics. Velocities for 14 layers in the model were derived from the analysis of five nearby VSPs (vP =1600–2300 m s-1). Synthetic shot data were compared with the field data and seismic sections were created. Major features like direct wave and reflections are imaged. We reproduce the mayor reflectors in the depth section of the field data, e.g. a prominent till layer and several deep reflectors. The SH-wave model was adapted accordingly but only led to minor correlation with the field data and produced a higher signal-to-noise ratio. Therefore, we suggest to consider for future simulations additional features like intrinsic damping, thin layering, or a near-surface weathering layer. These may lead to a better understanding of key parameters determining the data quality of near-surface shear-wave seismic measurements.

Surface wave excitation at the northern edge of the Kanto Basin, Japan

In this paper, we analyze the seismograms of a local earthquake (M W = 5.8) and find a clear surface wave excitation at the northern edge of the Kanto Basin, Japan. To conduct three-dimensional (3-D) finite difference method simulations for this observation, we investigate the characteristics of the S wave velocity in the Kanto Basin and construct a realistic sedimentary structure model. The S wave velocities derived from the vertical seismic profiling measurements in the Kanto Basin show smooth depth gradients rather than step-like increases and can be successfully modeled by an exponential asymptotically bounded velocity function. Our simulation using the modified Japan integrated velocity structure model by incorporating a sedimentary layer with S wave velocity-depth gradients well reproduces the excitation and dispersion of the observed surface waves. This result indicates that realistic modeling of the impedance contrast at the sediment-bedrock interface is indispensable for the precise evaluation of surface wave excitation at the basin edge.

Reliability of the slowness and slowness-polarization methods for anisotropy estimation in VTI media from 3C walkaway VSP data

We studied the validity of qP-wave slowness and slowness-polarization methods for estimating local anisotropy parameters in transversely isotropic (TI) media by quantifying the estimation errors in a numerical exercise. We generated numerical slownesses and polarizations over two aperture ranges corresponding to a short offset walkaway vertical seismic profiling (VSP) and a long offset walkaway VSP for a range of TI models with vertical axis of symmetry (VTI). Synthetic data are equisampled over the phase angle range and contaminated with Gaussian noise. We inverted the data and compared the anisotropy parameters of the optimal model with the true model. We found that the selection of a proper methodology for VTI parameter estimation based on walkaway VSP measurements was mostly dependent on our ability to accurately estimate either horizontal components of qP-wave slowness vector or the polarization vector. With data contaminated with noise, methods that include the horizontal component of the slowness vector had greater accuracy than the methods that replace this information with polarization measurements. The estimations are particularly accurate when a wide range of propagation angle was available. For short offsets, only parameter [Formula: see text] could be reliably estimated. In the absence of long offsets, depending on the accuracy of polarization measurements, the method based on the weak anisotropy approximation for qP-wave velocity in VTI media or the method based on slowness and polarization vectors could be used to estimate [Formula: see text] and [Formula: see text]. If the horizontal components of the slowness vector were not available (a heterogeneous overburden), we used methods that were based on local measurements of the polarization vector. We found that, with accurate measurements of the polarization vector, the method based on exact relationship between vertical slowness and polarization dip could be used to estimate VTI parameters even for the cases in which the wide offset range was not available.

Seismic anisotropy in the crystalline upper crust: observations and modelling from the Outokumpu scientific borehole, Finland

SUMMARY Seismic velocity anisotropy measurements are made of a fractured metamorphic formation from the 2.5-km-deep International Continental Scientific Drilling Programme (ICDP) borehole in Outokumpu, Finland. Three component walk-away vertical seismic profile (VSP) measurements are made along two source-line azimuths at three receiver depths (1000, 1750 and 2500 m) and incidence angle-dependent qP- and qS-wave velocities are extracted with a τ–p method. The highest qP-wave anisotropy, 13.6 per cent (vfast= 6160 m s−1, vslow= 5370 m s−1), is measured between 1000- and 1750-m depth, with anisotropy of up to 9.4 per cent (vfast= 6090 m s−1, vslow= 5540 m s−1) measured between 1750 and 2500 m depth. The top ∼1300 m of the region is composed of a homogeneous, strongly intrinsically anisotropic biotite-rich schist, and is sampled by the shallowest walk-away profile. Anisotropy of up to 11.1 per cent (vfast= 5950 m s−1, vslow= 5320 m s−1) is measured by the walk-away VSP between 50 and 1000 m depth, along with shear wave splitting averaging 5 per cent (180 m s−1). Laboratory-derived intrinsic anisotropy of the schist cannot by itself explain the degree or orientation of the anisotropy measured in the walk-away VSPs, however, a model which modifies the intrinsic stiffnesses by the inclusion of a single set of dipping, aligned cracks allows the observed in situ velocities to be reproduced. Forward modelling of the qP-wave walk-away VSP measurements from 50 to 1000 m depth is undertaken using an effective medium model to develop a 3-D velocity model of this region. An orthorhombic medium is used to represent the intrinsic anisotropy of the biotite-rich schist, and a single set of aligned cracks is added to give a bulk elastic stiffness. The resulting model predicts the schist to have an overall anisotropy of 16.8 per cent, with qP-wave velocities of up to 6315 m s−1. The accuracy of the model is assessed through its fit to the walk-away VSP measurements as well as a comparison to known geology of the region.

Hydrological aspects of geophysical borehole measurements in crystalline rocks of the Black Forest

One of six wells drilled into the crystalline rock of the Black Forest basement during the German Continental Deep Drilling Project (KTB) site survey has been used to detect hydraulically active fractures and to identify their character by a variety of logging tools such as caliper and televiewer logging, vertical seismic profiling (VSP) and electrical and thermal measurements in addition to standard logging recordings. In the 265 m Moosengrund well drilled into the Triberg granite body these methods were applied successfully to discriminate between fractures and breakouts and to identify open and closed fractures and water-consuming or water-expelling fractures, respectively. Fracture strike, dip and width were determined from televiewer data. VSP measurements were used to determine the seismic velocities in the unfractured rock and the fracture zones. Estimates of permeability (10–55 Darcy) and transmissivity (0.1–1.0cm 2s −1) of the fracture zones were obtained by the analysis of P waves and tube waves in the VSP records. The summed values of the transmissivity yield an upper estimate of the overall transmissivity determined by a pumping test.

Vertical seismic profile synthetics by dynamic ray tracing in laterally varying layered anisotropic structures

Dynamic ray tracing (DRT) is important in evaluating high‐frequency seismic wave fields in complicated structures. Two formulations of the DRT equations for inhomogeneous anisotropic media are presented. One of them is represented by the classical DRT system, suggested by Červený 15 years ago. Both systems are specified in Cartesian coordinates. The DRT equations are supplemented with formulae for the transformation of DRT across a smoothly curved interface between two inhomogeneous anisotropic media. Červený's formulation of the DRT is applied to the computation of vertical seismic profile (VSP) synthetics. The results of DRT are used for the evaluation of the geometrical spreading and of the coefficients of the paraxial ray approximation for travel times and ray amplitudes. In addition, the DRT is also used in the interval ray tracing procedure, a procedure searching for rays starling from the source and terminating in a specified interval on a profile. Results of numerical modeling of VSP measurements in a three‐dimensional laterally varying structure consisting of Isotropic and anisotropic layers separated by curved interfaces are presented. Ray diagrams of selected elementary waves, time‐distance curves, and multisource three‐component VSP synthetics generated for two different source locations calculated for the anisotropic model are compared with the results determined for a reference Isotropic model. The latter is obtained by averaging phase velocities of the anisotropic model. This comparison clearly shows the effects of anisotropy and the lateral variation of the model on seismic wave fields. The reliability of the ray synthetics is briefly discussed.

The use of an active drill bit for inverse VSP measurements

Vertical Seismic Profiling (VSP) is often used to provide high resolution seismic images near a wellbore. A new borehole seismic technique, the TOMEX� survey (Rector, et al., 1988), uses the vibrations produced by a drill bit as a downhole seismic energy source to produce inverse VSP data. No downhole instrumentation is required to acquire the data, and the data recording does not interfere with or delay the drilling process. Hence, there is no loss of rig time in performing the survey. These characteristics offer a method to acquire SWD (seismic-while-drilling) borehole seismic surveys. In addition, 3-D imaging around a well can be obtained at significant savings compared to conventional offset VSP imaging. The continuous signals generated by the bit during drilling are monitored with a reference sensor attached to the top of the drillstring, and the reference sensor signals are crosscorrelated with signals from surface-positioned geophones to produce inverse VSP data. Deconvolution and time shifts are then performed to remove the effects of recording the source reference trace at a location that is a considerable distance from the source. Results from tests demonstrate that these processed drill-bit source data are virtually identical to conventional forward VSP data. In using the drill bit as a downhole seismic source for inverse VSP, many of the limitations of conventional VSP are overcome. Several applications for VSP that were previously considered to be prohibitively expensive are now feasible. Furthermore, this seismic-while-drilling technique offers the potential for the explorationist to make real-time drilling decisions at the well site.

Particle displacements on the wall of a borehole from incident plane waves

Particle displacements from incident plane waves at the wall of a fluid‐filled borehole are formulated by applying the seismic reciprocity theorem to far‐field displacement fields. Such displacement fields are due to point forces acting on a fluid‐filled borehole under the assumption of long wavelengths. The displacement fields are analyzed to examine the effect of the borehole on seismic wave propagation, particularly for vertical seismic profiling (VSP) measurements. When the shortest wavelength of interest is approximately 25 times longer than the borehole’s diameter, the scattered displacements are proportional to the first power of incident frequency and borehole diameter. The maximum scattered energy occurs when an incident P‐wave propagates perpendicular to the borehole. Borehole effects on VSP measurements, such as waveform distortion, amplitude variation, and time delay, have been analyzed using the concept of a transfer function. When the shortest wavelength of interest is about 40 times longer than the borehole’s diameter, borehole effects on VSP measurements using a wall‐locking geophone are negligible.

Plane‐wave Q deconvolution

Often the information content of measured signals from distance sources is hidden, because the signal distorts, weakens, and loses resolution as it propagates. For seismic energy traveling in the earth, these propagation effects can be approximated by the constant (frequency‐independent) Q model for attenuation and dispersion. For a propagating plane wave, this model leads to a spatial attenuation factor that is an unbounded function of frequency. Consequently, the broadband inverse of the constant-Q filter does not exist. For a fixed distance between the source and receiver the effects of the propagation path can be deconvolved (removed) within the seismic band by reversing the propagation of the plane wave. This propagation reversal is done by a time reversal with Q replaced by —Q, thereby changing absorption to gain in the complex wavenumber. Normally, measured seismic traces contain returns from a variety of depths. The interference of waves with different amounts of attenuation complicates the inversion process. From a superposition of plane waves with reversed propagation, a general inverse to an attenuation earth filter is proposed. To account for the increased attenuation with depth, the plane‐wave inverse filter is now time‐varying. This time‐varying inverse filter has a simple Fourier integral representation where the wavenumber is complex, and the direction of propagation is chosen such that the wave is growing rather than attenuating with distance. To control the wavelet side lobes a frequency‐domain window function (Hanning window) is applied to the trace. This two‐step plane‐wave deconvolution scheme was demonstrated to be superior to conventional deconvolution procedures. Tests with field data indicate the method is effective in removing attenuation effects from both VSP (Vertical Seismic Profile) and surface measurements. Phase distortions are eliminated and interference between events is reduced within the seismic band. This inverse is nearly exact for events where the time‐bandwidth (propagation time‐signal bandwidth) product is less than the effective Q. For depths where the time‐bandwidth product is greater than [Formula: see text] large wavelet side lobes appear. The wavelet side lobes can be partially suppressed by tapering the edges of the spectrum. However, the large side lobes of wavelets from shallow reflectors limit the bandwidth that can be recovered from the deeper events to aproximately [Formula: see text], where t is the propagation time to the event. Advances in the inversion algorithm (e.g., a Wiener filter could be used in place of the Hanning window to control side lobes) could probably improve upon our results, but in most cases even a small amount of measurement noise limits the reflection sequences to time‐bandwidth products that are less than twice the effective Q.