Time Migration Velocity Research Articles

A new strategy for prestack time migration velocity analysis to assess and mitigate structural uncertainty

Velocity uncertainty represents a significant contributor to structural uncertainty in seismic imaging. Therefore, improving velocities and enhancing the resolution of velocity spectra are crucial for reducing structural uncertainty. In the present study, a novel approach is introduced to increase velocity resolution and mitigate structural uncertainty. The methodology's impact on spatial and temporal uncertainty of image points is compared with the improved velocity continuation method. The proposed strategy involves modifying the equations of the weighted semblance method and integrating them into the equations of the velocity continuation method. Specifically, the vertical analysis window of the weighted semblance, typically applied on coherency panels is changed to a horizontal window that is applied on common-image gathers. Correspondingly, the weighting factor is adjusted using a time-adaptive function. The velocity uncertainty equation in the velocity continuation method is then replaced by the new weighted semblance eq. A similar substitution can be performed for structural uncertainty equations. These new equations are then used on common-image gathers obtained by the velocity continuation method to increase the accuracy of the velocity model and then investigate changes in structural uncertainty. In the proposed strategy, prestack time migration is initially conducted on seismic common-offset sections using various constant velocities within a predefined range. Subsequently, the weighted semblance technique is used for automatic velocity model updating in all common-image gathers. The new velocity model is then used for final prestack time migration. In the equations pertaining to velocity uncertainty, a width equivalent to one standard deviation on both sides of the picked velocity value is considered as the uncertainty range. By narrowing this width through the proposed method, it is expected to mitigate subsequent structural uncertainty, specifically for 1D media. Results from applying the presented method on synthetic and marine seismic data demonstrate its effectiveness in reducing structural uncertainty for 1D time migrated images compared to the velocity continuation method. It is also shown that horizontal structural uncertainty is significantly more sensitive to velocity errors rather than vertical uncertainty. An added benefit is that the increased resolution in velocity spectra also provides appropriate panels for automatic velocity picking.

Efficient prestack time migration velocity analysis: A correlation‐based global optimization approach

AbstractAn accurate migration velocity model is vital to seismic imaging. Conventional time‐domain migration velocity analysis techniques commonly obtain the velocity model by iteratively picking the velocity spectra of common reflection point gathers or scanning selection, which is usually laborious and computationally expensive. To solve these issues, we propose a more effective migration velocity analysis method for prestack time migration. This method is based on cross‐correlation stacked time shift functions of local migrated gathers and uses the very fast simulated annealing algorithm to semi‐automatically invert the migration velocity parameters. This new correlation‐based objective function can catch subtle bending features of the reflector in the local migrated gather. The proposed approach applies an efficient global optimization algorithm that does not depend on the initial parameter value and is easy to extend to the multiparameter complex case. Moreover, the strategy of using local migrated gathers can incorporate prior information in the inversion to avoid the effects of misidentifying reflection events and the interference of multiples. Furthermore, applying a dynamic parameter constraint strategy for the very fast simulated annealing algorithm improves the convergence speed. Both synthetic and real data examples demonstrate that the proposed migration velocity analysis method can efficiently build an accurate time migration velocity model, which can also provide a good initial velocity model for depth migration.

Reprocessing 2-D airgun seismic reflection data SALTFLU (salt deformation and sub-salt fluid circulation in the Algero-Balearic abyssal plain) in the Balearic promontory and the Algerian basin

In an ever more challenging context for the acquisition of seismic data in the Mediterranean Sea, reprocessing to improve the quality of legacy data has become increasingly important. This work presents the newly reprocessed, open access dataset SALTFLU acquired in the Algerian basin by the National Institute of Oceanography and Applied Geophysics (OGS) in 2012. We apply a ‘broadband’ reprocessing strategy adapted for offset-limited (3 km streamer for a target 4 km below the sea level) airgun reflection seismic data acquired in deep water settings. We then assess if the reprocessed images provide new geological insights on the Mediterranean sub-surface. The workflow relies on an integrated approach combining geophysics and geological interpretation to iteratively build the velocity model. In this way we aim to tackle some of the challenges linked to imaging deep complex geological structures containing high velocity contrasts with 2-D, offset-limited seismic data. We first broaden the bandwidth of the data through multi-domain de-noising, deghosting and a source designature using an operator derived from the seabed reflection. We then perform iterative migration velocity analysis, pre-stack time migration and multiple attenuation in the Radon domain to obtain time-migrated images. The initial velocity model is derived from the resulting time migration velocities, and geologically driven model updates are generated using a combination of travel-time tomography, seismic interpretation of the major salt horizons and velocity gradient flooding. The gradient flooding aims to reproduce the large scale first-order velocity variations, while the travel-time tomography aims to resolve the smaller second-order velocity variations. The results improve our deep geological knowledge of the under-explored Algerian basin down to the base salt and the pre-salt. Fluid indicators are imaged within the Plio-Quaternary of the Algerian basin, which we interpret as thermogenic or biogenic gas sourced from either the Messinian Upper Unit or from the pre-salt, migrating through a hydro-fractured salt. The reprocessed data image lateral and vertical seismic facies variation within the Messinian units that could shed new light on the tectono-stratigraphic processes acting during the Messinian Salinity Crisis. It also reveals numerous previously unresolved volcanic structures within the Formentera basin.

Migration velocity analysis based on focusing diffractions generated from CRS wavefront attributes: Application to real land data

In land seismic data processing, the prestack time migration (PSTM) image remains the standard imaging output, but a reliable migrated image of the subsurface depends on the accuracy of the migration velocity model. We have adopted two new algorithms for time-domain migration velocity analysis based on wavefield attributes of the common-reflection-surface (CRS) stack method. These attributes, extracted from multicoverage data, were successfully applied to build the velocity model in the depth domain through tomographic inversion of the normal-incidence-point (NIP) wave. However, there is no practical and reliable method for determining an accurate and geologically consistent time-migration velocity model from these CRS attributes. We introduce an interactive method to determine the migration velocity model in the time domain based on the application of NIP wave attributes and the CRS stacking operator for diffractions, to generate synthetic diffractions on the reflection events of the zero-offset (ZO) CRS stacked section. In the ZO data with diffractions, the poststack time migration is applied with a set of constant velocities, and the migration velocities are then selected through a focusing analysis of the simulated diffractions. We also introduce an algorithm to automatically calculate the migration velocity model from the CRS attributes picked for the main reflection events in the ZO data. We determine the precision of our diffraction focusing velocity analysis and the automatic velocity calculation algorithms using two synthetic models. We also applied them to real 2D land data with low quality and low fold to estimate the time-domain migration velocity model. The velocity models obtained through our methods were validated by applying them in the Kirchhoff PSTM of real data, in which the velocity model from the diffraction focusing analysis provided significant improvements in the quality of the migrated image compared to the legacy image and to the migrated image obtained using the automatically calculated velocity model.

High-Resolution Velocity Model Building Based on Common-Source Migration Images and Convolutional Neural Networks

Building a reliable velocity model plays a vital role in seismic imaging and quantitative reservoir description. However, the current data-driven deep-learning-based velocity model building (VMB) approaches directly reconstruct the velocity model of the subsurface from prestack seismic recordings, which are very sensitive to the noise and amplitude mismatch in the data domain. In this letter, we propose a novel VMB approach based on common-source migration image gathers (CSMIGs) and convolutional neural networks (CNNs). The proposed CNN architecture uses the CSMIGs reconstructed by the reverse time migration approach and migration velocity model as the input data. It aims to capture the nonlinear relationship between the amplitude and phase information of CSMIGs and the optimal subsurface reflectivity model. Trained with realistic subsurface models, it can determine that the VMB approach is a computationally efficient solution for a high-resolution velocity reconstruction. In addition, the proposed approach has a better reconstruction performance, antinoise ability, and can be generalized much more easily than the data-driven VMB approach.

Improving prestack time migration by least-squares equivalent offset method

Time migration is an important tool to provide a subsurface image because it is faster and less sensitive to velocity errors than depth migration. However, a reliable and focused time migration image is achievable only with well-determined time-migration velocities. As an alternative of prestack time migration, equivalent offset migration (EOM) provides intermediate common scatterpoint (CSP) gathers that have higher fold and larger offset range than common midpoint (CMP) gathers, which enables precise focusing of the velocity semblance and accurate velocity analysis. However, direct implementation of EOM might introduce coherent mapping noise, which smears the velocity spectra and decreases the resolution of the final stacked image. We reformulate the EOM as a linear operator that maps CMP gathers to CSP gathers and then propose least-squares equivalent offset migration (LS-EOM) method that inverts for optimal CSP gathers by minimising the misfit between simulated and observed CMP gathers. Compared with the EOM method, LS-EOM provides noise-free CSP gathers, leading to better focusing of the velocity semblance and higher resolution of the final stacked image. Both synthetic and field data tests demonstrate the effectiveness of our proposed method.

Prestack time migration velocity analysis using recurrent neural networks

We present a new efficient method to perform prestack time migration velocity analysis (MVA) of seismic data based on recurrent neural nets (RNNs). We assume that there exits a mapping from each local data subvolume, which we term an analysis volume, to each single velocity point value in space. Under this hypothesis, as RNNs are capable of learning structural information both in time and in space, we propose designing a net that can learn this mapping, and exploit it to produce the entire root-mean-square (rms) velocity field. The performance of the method is evaluated via real data experimental results in comparison to existing technique for MVA. We present different aspects of the system's behavior with various training and testing scenarios and discuss potential advantages and disadvantages. We also recast the same RNN to learn and automate migration velocity picking from constant velocity migration (CVM) panels and compare real data results for this alternative. Our approach is extremely efficient in terms of computational complexity and running time, and therefore can be potentially applied to large volumes of three-dimensional (3D) seismic data, and significantly reduce work load. An extension to four-dimensional (4D) data is also possible.

Time-migration velocity estimation using Fréchet derivatives based on nonlinear kinematic migration/demigration solvers

Advanced seismic imaging and inversion are dependent on a velocity model that is sufficiently accurate to render reliable and meaningful results. For that reason, methods for extracting such velocity models from seismic data are always in high demand and are topics of active investigation. Velocity models can be obtained from both the time and depth domains. Relying on the former, time migration is an inexpensive, quick and robust process. In spite of its limitations, especially in the case of complex geologies, time migration can, in many instances (e.g. simple to moderate geological structures), produce image results compatible to the those required for the project at hand. An accurate time-velocity model can be of great use in the construction of an initial depth-velocity model, from which a high-quality depth image can be produced. Based on available explicit and analytical expressions that relate the kinematic attributes (namely, traveltimes and local slopes) of local events in the recording (demigration) and migrated domains, we revisit tomographic methodologies for velocity-model building, with a specific focus on the time domain, and on those that makes use of local slopes, as well as traveltimes, as key attributes for imaging. We also adopt the strategy of estimating local inclinations in the time-migrated domain (where we have less noise and better focus) and use demigration to estimate those inclinations in the recording domain. On the theoretical side, the main contributions of this work are twofold: 1) we base the velocity model estimation on kinematic migration/demigration techniques that are nonlinear (and therefore more accurate than simplistic linear approaches) and 2) the corresponding Frechet derivatives take into account that the velocity model is laterally heterogeneous. In addition to providing the comprehensive mathematical algorithms involved, three proof-of-concept numerical examples are demonstrated, which confirm the potential of our methodology.

Imaging zero-offset 3-D P-cable data with CRS method

SUMMARYStandard seismic acquisition and processing require appropriate source–receiver offsets. P-cable technology represents the opposite, namely, very short source–receiver offsets at the price of increased spatial and lateral resolution with a high-frequency source. To use this advantage, a processing flow excluding offset information is required. This aim can be achieved with a processing tuned to diffractions because point diffractions scatter the same information in the offset and midpoint direction. Usually, diffractions are small amplitude events and a careful diffraction separation is required as a first step. We suggest the strategy to use a multiparameter stacking operator, for example, common-reflection surface, and stack along the midpoint direction. The obtained kinematic wave-front attributes are used to calculate time-migration velocities. A diffractivity map serves as a filter to refine the velocities. This strategy is applied to a 3-D P-cable data set to obtain a time-migrated image.

Effects of lateral heterogeneity on time-domain processing parameters

SUMMARY Time-domain processing of seismic reflection data has always been an important engine that is routinely utilized to produce seismic images and to expeditiously construct subsurface models. The conventional procedure involves analysing parameters related to the derivatives of reflection traveltime with respect to offset including normal moveout (NMO) velocities (second-order derivatives) and quartic coefficients (fourth-order derivatives). In this study, we propose to go beyond the typical assumption of 1-D laterally homogeneous medium when relating those ‘processing’ parameters to the subsurface medium parameters and take into account the additional influences from lateral heterogeneity including curved interfaces and smoothly variable velocities. We fill in the theoretical gap from previous studies and develop a general framework for such connection in layered anisotropic media. We show that in general, the influences of lateral heterogeneity get accumulated from all layers via a recursive relationship according to the Fermat’s principle and can be approximately quantified in terms of the lateral derivatives of the layer interface surfaces and velocities. Based on the same general principle, we show that our approach can also be used to study the lateral heterogeneity effects on diffraction traveltime and its second-order derivative related to time-migration velocity. In this paper, we explicitly specify expressions for NMO and time-migration velocities with the influences from both types of heterogeneity suitable for 2-D data sets and also discuss possible extensions of the proposed theory to 3-D data sets and to parameters related to higher-order traveltime derivatives. Using numerical examples, we demonstrate that the proposed theory can lead to more accurate reflection and diffraction traveltime predictions in comparison with those obtained based on the 1-D assumption. Both the proposed theoretical framework and its numerical testing for forward traveltime computation presented in this study aid in understanding the effects from lateral heterogeneity on time-processing parameters and also serve as an important basis for designing an efficient technique to separate those influences in important processes such as Dix inversion for a more accurate subsurface model in the future.

Velocity-estimation improvements and migration/demigration using the common-reflection surface with continuing deconvolution in the time domain

To obtain an image of the earth’s subsurface, time-imaging methods can be applied because they are reasonably fast, are less sensitive to velocity model errors than depth-imaging methods, and are usually easy to parallelize. A powerful tool for time imaging consists of a series of prestack time migrations and demigrations. We have applied multiparameter stacking techniques to obtain an initial time-migration velocity model. The velocity model building proposed here is based on the kinematic wavefield attributes of the common-reflection surface (CRS) method. A subsequent refinement of the velocities uses a coherence filter that is based on a predetermined threshold, followed by an interpolation and smoothing. Then, we perform a migration deconvolution to obtain the final time-migrated image. The migration deconvolution consists of one iteration of least-squares migration with an estimated Hessian. We estimate the Hessian by nonstationary matching filters, i.e., in a data-driven fashion. The model building uses the framework of the CRS, and the migration deconvolution is fully automated. Therefore, minimal user interaction is required to carry out the velocity model refinement and the image update. We apply the velocity refinement and migration deconvolution approaches to complex synthetic and field data.

Coherence-based time migration velocity analysis by the use of supergathers

Errors in the migration velocity field are perceived as moveouts in common-image gathers (CIGs). These moveouts depend on discrepancies in the velocity field as well as in the dip of the reflectors. We extend previous works into two main aspects. First, we found an accurate description of the residual moveout, considering the reflector dip, whereas only second-order approximations have been provided so far. Our formula for the moveout is exact, at the cost of being implicit. We evaluate an efficient numerical method to compute it accurately. Examples are provided to confirm that dipping angles far beyond those considered by previous approaches are reliable. Second, our moveout description is valid within a range of offsets and CIG positions, meaning that we can handle more than one CIG at once. In other words, instead of describing a moveout curve on a single CIG, we are able to compute a moveout surface across image gathers. From this larger volume of data, numerical experiments determine that more robust estimations are obtained for the velocity correction factors. Finally, the estimation of the best-fit parameters is carried out by a traditional coherence-based analysis, in which a modern derivative-free optimization method is used. We develop a set of numerical tests to validate the description of the residual moveouts, to compare our strategy with previous work from the literature, as well as to demonstrate that just a few image gathers are sufficient to stabilize the estimation of velocity correction factors. As a by-product, we obtained estimations of the reflector as well.

Common-reflection-surface-based prestack diffraction separation and imaging

Diffraction imaging can lead to high-resolution characterization of small-scale subsurface structures. A key step of diffraction imaging and tomography is diffraction separation and enhancement, especially in the full prestack data volume. We have considered point diffractors and developed a robust and fully data-driven workflow for prestack diffraction separation based on wavefront attributes, which are determined using the common-reflection-surface (CRS) method. In the first of two steps, we apply a zero-offset-based extrapolation operator for prestack diffraction separation, which combines the robustness and stability of the zero-offset CRS processing with enhanced resolution and improved illumination of the finite-offset CRS processing. In the second step, when the finite-offset diffracted events are separated, we apply a diffraction-based time migration velocity model building that provides high-quality diffraction velocity spectra. Applications of the new workflow to 2D/3D complex synthetic data confirm the superiority of prestack diffraction separation over the poststack method as well as the high potential of diffractions for improved time imaging.

Diffraction imaging and time-migration velocity analysis using oriented velocity continuation

We perform seismic diffraction imaging and time-migration velocity analysis by separating diffractions from specular reflections and decomposing them into slope components. We image the slope components using migration velocity extrapolation in time-space-slope coordinates. The extrapolation is described by a convection-type partial differential equation and implemented in a highly parallel manner in the Fourier domain. Synthetic and field data experiments show that the proposed algorithms are able to detect accurate time-migration velocities by measuring the flatness of diffraction events in slope gathers for single- and multiple-offset data.

Extracting geologic information directly from high-resolution full-waveform inversion velocity models — A case study from offshore Trinidad

The East Coast Marine Area (ECMA) of Trinidad and Tobago contains producing gas fields where the imaging of seismic data is known to be challenging. This difficulty has been attributed to the combination of the Dolphin Main Fault, with more than 1 km of throw and large velocity contrasts of up to 30% in magnitude across it, and shallow gas and gas clouds causing very high attenuation. These factors lead to difficulties producing reliable and consistent velocity models. Previous velocity models — generated using traveltime tomography — contained inaccuracies that led to poor imaging and structural positioning, resulting in uncertainty when planning exploration and development drilling programs. In an attempt to reduce these uncertainties, a full-waveform inversion (FWI) feasibility study was performed over the most problematic area, with an initial velocity model built from a combination of legacy prestack time and depth migration velocity models and additional first-break refraction tomography. The geologic complexity near the main fault and the associated poor seismic data quality meant this model gave a poor fit between the synthetic and observed data (from a cycle-skipping point of view) at the lowest usable frequency. This led to the idea of applying a diving-wave-driven, offset-stripping workflow across each frequency band in the FWI, going from 3.5 Hz to 7.5 Hz. After successful application on the test area, the workflow was applied to ∼1400 km2 of 3D seismic over the ECMA. The final velocity model is structurally consistent and highly resolved, containing detailed shallow geologic features such as mud volcanoes, well-defined faults, and potential gas anomalies that were not visible on the legacy tomographic model. A blind test of well-pressure data correlates strongly with velocity drops in the FWI model, showing that FWI models can potentially provide very good input to regional pore-pressure-prediction studies.

The offset-midpoint traveltime pyramid of P-waves in homogeneous orthorhombic media

The offset-midpoint traveltime pyramid describes the diffraction traveltime of a point diffractor in homogeneous media. We have developed an analytic approximation for the P-wave offset-midpoint traveltime pyramid for homogeneous orthorhombic media. In this approximation, a perturbation method and the Shanks transform were implemented to derive the analytic expressions for the horizontal slowness components of P-waves in orthorhombic media. Numerical examples were shown to analyze the proposed traveltime pyramid formula and determined its accuracy and the application in calculating migration isochrones and reflection traveltime. The proposed offset-midpoint traveltime formula is useful for Kirchhoff prestack time migration and migration velocity analysis for orthorhombic media.

Automatic time-domain velocity estimation based on an accelerated clustering method

Time-domain velocity and moveout parameters can be directly obtained from local event slopes, which are estimated on the prestack seismic gathers. In practice, there are always some errors in the estimated local slopes, especially in low signal-to-noise ratio (S/N) situations. Thus, subsurface velocity information may be hidden in the image domain spanned by velocity and other moveout parameters. We have developed an accelerated clustering algorithm to find cluster centers without prior information about the number of clusters. First, plane-wave destruction is implemented to estimate the local event slopes. For every sample in the seismic gathers, we obtain the estimations of velocity and its location in the image domain, according to the local event slopes. These mapped data points in the new domain exhibit the structure of groups. We represent these points by a mixture distribution model. Then, the cluster centers of the mixture distribution model are located, which correspond to maximum likelihood velocities of the main subsurface structures. Approximate velocity uncertainties bounds are used to select centers corresponding to reflections. Finally, interpolation is performed on the clustered unevenly sampled knot velocities to build the effective velocity model on regular grids. With synthetic and field data examples, we have determined that the proposed automatic velocity estimation method can give a stacking velocity model and a time migration velocity model with relatively high accuracy.

Prestack migration velocity analysis based on simplified two-parameter moveout equation

Stacking velocity VC2, vertical velocity ratio γ0, effective velocity ratio γeff, and anisotropic parameter χeff are correlated in the PS-converted-wave (PS-wave) anisotropic prestack Kirchhoff time migration (PKTM) velocity model and are thus difficult to independently determine. We extended the simplified two-parameter (stacking velocity VC2 and anisotropic parameter keff) moveout equation from stacking velocity analysis to PKTM velocity model updating and formed a new four-parameter (stacking velocity VC2, vertical velocity ratio γ0, effective velocity ratio γeff, and anisotropic parameter keff) PS-wave anisotropic PKTM velocity model updating and process flow based on the simplified two-parameter moveout equation. In the proposed method, first, the PS-wave two-parameter stacking velocity is analyzed to obtain the anisotropic PKTM initial velocity and anisotropic parameters; then, the velocity and anisotropic parameters are corrected by analyzing the residual moveout on common imaging point gathers after prestack time migration. The vertical velocity ratio γ0 of the prestack time migration velocity model is obtained with an appropriate method utilizing the P- and PS-wave stacked sections after level calibration. The initial effective velocity ratio γeff is calculated using the Thomsen (1999) equation in combination with the P-wave velocity analysis; ultimately, the final velocity model of the effective velocity ratio γeff is obtained by percentage scanning migration. This method simplifies the PS-wave parameter estimation in high-quality imaging, reduces the uncertainty of multiparameter estimations, and obtains good imaging results in practice.

Prestack time-migration velocity analysis using remigration trajectories

We have developed a prestack time-migration tool for local improvement of the seismic migration-velocity model. The method is based on remigration trajectories that describe the position of an image point in the image domain for different source-receiver offsets as a function of the migration velocity. It determines kinematic migration parameters using local-slope information of migrated seismic reflection events. These parameters, in turn, are used to locally correct the velocity model. The main advantage of this technique is that it allows us to carry out a moveout correction not just at a fixed point in a zero-offset (poststack) time-migrated gather, but varying through all offsets of a common-image gather, taking into account the reflection-point displacement in the midpoint direction. In other words, it provides for time-migration velocity analysis from prestack data. We have tested the feasibility of the method on synthetic data from four simple models and the Sigsbee2B data. Our tests determined that the proposed tool increases the velocity-model resolution and provides a plausible time-migrated image. The quality of the initial model is not critical. The procedure is quite efficient. Significant effort was spent on manual event picking.

Migration-velocity building in time and depth from 3D (2D) Common-Reflection-Surface (CRS) stacking - theoretical framework

We demonstrate how the velocity concepts attached to conventional stacking assuming local smooth 1D type of Earth models, are modified within the setting of the Common Reflection Surface (CRS) method to handle lateral velocity variations. The corresponding matrix (scalar) normal moveout (NMO)- and Dix-velocities in 3D (2D) are now linked to a smooth velocity medium in depth sampled along the mapping or normal rays taking into account lateral velocities. This is rather different from the conventional 1D case where the link between the data-driven velocities and the smooth 1D velocity medium are represented by vertical mapping rays. Further and by analogy with the conventional 1D approach, where time-migration velocities are computed from NMO- or Dix-velocities after proper smoothing, its CRS counterpart is established. It is demonstrated that matrix (scalar) time-migration velocities to be used in ray-based 3D (2D) time-migration (TM) can be obtained by proper mapping of the corresponding CRS velocities. Relevant mapping equations valid for the 2D case have earlier been treated in the literature. However, to our knowledge, this is the first time such mapping equations have been derived for the 3D case. This mapping takes into account smooth lateral velocity variations, normally not properly accounted for in the initial model employed in conventional time-migration building based on NMO-velocities. Also, unlike the conventional approach, the use of the CRS method makes it feasible to build a smooth velocity field in depth beyond that of 1D. In this connection, we propose to use Normal Incidence Point (NIP) tomography (alternatively Image Incidence Point (IIP) tomography) to build a macro-velocity model upon which the time-to-depth mapping and the correction factors can be calculated. That step is then followed by construction of an updated smooth replacement medium, better able to capture local changes in velocity. The proposed approach can be used iteratively. Such depth velocities can be employed as an initial model for iterative depth migration. The main idea being that velocity-model building by using time-domain observables ensures more robustness.