Abstract: Method for building a subsalt velocity model. In one implementation, the method may include determining a velocity of a sedimentary area surrounding a salt body, determining an initial velocity of a subsalt sedimentary area disposed below the salt body by interpolating the velocity of the surrounding sedimentary area with one or more boundaries defining the salt body, and updating the initial velocity of the subsalt sedimentary area.

Abstract: A technique includes modeling a seismic wavefield in a vertical transversely isotropic media as a second order derivative of a first wavefield function with respect to a vertical direction and not with respect to crossline and inline directions and as second order derivatives of a second wave field function with respect to the inline and crossline directions and not with respect to the vertical direction. The method includes processing seismic measurements of the media in accordance with the modeled seismic wavefield to obtain information about the media.

Abstract: Correcting seismograms to compensate for absorption or dissipative effects that occur in the earth. In one implementation, a ratio of traveltime (t) to absorption parameter (Q) may be computed for each seismogram. The ratio may be referred to as R. Further, the ratio may be computed for a predetermined set of traveltimes. As a result, a system of linear equations may be generated, where each ratio is represented by a linear equation having a number of unknown components. The system of linear equations may then be solved for the unknown components. The solved components may then be recombined or added to generate an estimate of R. In one implementation, only a portion of the solved components may be added. The seismograms may then be corrected using the estimate of R. As such, the seismograms may be corrected for absorption effects in a surface consistent manner.

Abstract: A method of predicting subsurface properties of a geologic formation includes acquiring seismic data for a subsurface region including the geologic formation of interest, computing seismic attributes of the measured seismic data over at least part of this geologic formation, determining internally consistent rock properties representative of the geologic formation, generating models of the same part of the geologic formation with these rock properties, computing synthetic seismic data from the models, computing the same attributes from these synthetic seismic data, and using Bayesian analysis to predict, from the probability of modeled attributes given the models, the probability of the actual subsurface properties given the measured attributes.

Abstract: To generate an absorption parameter model, estimated values of an effective absorption parameter are received, where the estimated effective absorption parameter values represent absorption encountered by a seismic wave in a subterranean structure. Based on the estimated effective absorption parameter values, an absorption parameter model is generated that varies absorption parameter values along at least one dimension of the subterranean structure.

Abstract: Method for updating a velocity model (926) for migrating seismic data using migration velocity scans with the objective of building a model that reproduces the same travel times that produced selected optimal images from a scan. For each optimal pick location (914) in the corresponding test velocity model (916), a corresponding location is determined (922) in the velocity model to be updated, using a criterion that the travel time to the surface for a zero offset ray (918) should be the same. Imaging travel times are then computed from the determined location to various surface locations in the update model (924), and those times are compared to travel times in the test velocity model from the optimal pick location to the same array of surface locations. The updating process consists of adjusting the model to minimize the travel time differences (934).

Abstract: According to a preferred aspect of the instant invention, there is provided herein a system and method for using both a temporal zone of seismic data and conditioned amplitude spectra thereof to determine the underlying layer sequence. The instant invention is suitable both for determining layer positions and layer impedance magnitudes, that best describes the geologic sequence sampled by a band limited seismic source wavelet as reflected from the zone.

Abstract: In a first aspect, a method for use in a time lapse, marine seismic survey includes accessing a set of baseline seismic data; accessing a set of acquired, time lapse seismic data; and interpolating a set of time lapse seismic data from the baseline seismic data and the acquired time lapse seismic data, at least one of the baseline seismic data and the acquired time lapse seismic data being multicomponent data. In other aspects, a program storage medium is encoded with instructions that, when executed by a computing device, perform the above method and a computing apparatus programmed to perform one or more of such methods.

Abstract: A technique is provided for identifying an internal multiple generator in a subterranean structure. The technique includes injecting wavefields at different levels in the subterranean structure, where the different levels are proximate a predicted location of the internal multiple generator. Wavefields induced by the injected wavefields are recorded and the effect of the internal multiple generator based on the recorded wavefields is determined.

Abstract: A method of constructing a 3D geologically plausible velocity model for efficient and accurate prestack imaging wherein embodiments of the invention provide: (1) a method of calibrating velocity functions, appropriately and effectively taking into account well (hard) and seismic (soft) data as well as geological features, and trend fitting (“iDEPTHing”) RMS velocities before curved-ray prestack time migration; (2) a method of calibrating and trend fitting (“iDEPTHing”) interval velocities before prestack depth migration, appropriately and effectively taking into account well (hard) and seismic (soft) data as well as geological features; and (3) a method of constructing a geologically plausible velocity model using the previous steps of velocity calibration and trend fitting RMS and interval velocities, for efficient sequential use in prestack time migration followed by prestack depth migration.

Abstract: A method for creating an enhanced seismic image is described. Seismic data is acquired from a seismic survey conducted over a subterranean region. The seismic data is transformed into energy components, preferably Gaussian beam components. An earth model is created which is comprised of lens elements. The set of energy components is propagated or migrated through the lens elements to form image components which are combined into a seismic image. A target is identified in the seismic image for image enhancement. Ray tracing may be used to select the trial set of lens elements to be updated and to select a subset of energy components. The subset of energy components is propagated through updated earth model to form updated image components. The seismic image is updated by replacing image components with updated image components which are formed from the subset of selected energy components. This subset is ideally greatly reduced in size relative to the overall number of energy components, i.e., beam components.

Abstract: Methods and systems for processing microseismic waveforms. The methods and systems provide determining a measure of waveform fit in the frequency-domain comprising constructing, in the frequency-domain, at least one of an amplitude misfit functional and a cross phase functional between arrivals; and estimating source parameters and/or model parameters.

Abstract: The invention is a method for defining the development conditions of a hydrocarbon reservoir by updating a velocity model by means of log data. A first seismic depth image representative of the reservoir is established by using seismic data and a velocity model. A series of log data is acquired from wells. Differences between seismic reflector depths observed in the first seismic depth image and depths for these reflectors identified in the wells are measured. The velocity model is modified to minimize these errors, using a prestack kinematic inversion technique allowing constraints to be taken into account. A new seismic depth image from which the development conditions of the hydrocarbon reservoir are determined is deduced therefrom. The invention has application for hydrocarbon reservoir development.

Abstract: A method of determining the existence of and location of hydrocarbon and water fluid contacts by analyzing spatial changes in 3D seismic data. The method begins by obtaining seismic attribute data as a 3D data volume. Then the method derives 3D dip and azimuth as a 3D volume and deriving corresponding 3D reliability volumes or derives responding 3D censor volumes which are representative of portions of the volume within which a reliable dip and azimuth can be determined. A focused subvolume of interest within the 3D data volume is then selected. The method concludes by storing filtered 3D seismic attribute data within the filter half length in each 3D direction along with null values for all non-reliability locations outside the filter half length but within the focused subvolume of interest forming a 3D output volume.

Abstract: The analysis of scattering diagrams of the correlation-type representation theorem in inhomogeneous media is improved by the use of virtual events. Virtual events here are events which are not directly recorded in standard seismic data acquisition, but the assumption of their existence permits the construction, of internal multiples with scattering points at the sea surface.

Abstract: A method of determining and analyzing spatial changes in the earth's subsurface. The method includes obtaining seismic attribute data as a 3D data volume and obtaining corresponding 3D dip and azimuth as a 3D volume and obtain corresponding 3D reliability volumes or 3D censor volumes which are representative of portions of the volume within in which a reliable dip and azimuth can be determined. The gradient of the seismic attribute data in the direction of structural dip is formed using either a dot product methodology or a derivative methodology after interpolation onto the direction of structural dip. At non-reliability locations or locations where no gradient could be meaningfully computed null values are stored. High gradient values in narrow time or depth ranges which are both statistically significant relative to a background level and contiguous designate regions likely to be proximal to a fluid contact or seismic flat spot.

Abstract: A method of determining and analyzing spatial changes in the earth's subsurface. The method includes obtaining seismic attribute data as a 3D data volume and obtaining corresponding 3D dip and azimuth as a 3D volume and obtain corresponding 3D reliability volumes or 3D censor volumes which are representative of portions of the volume within in which a reliable dip and azimuth can be determined. A focused subvolume of interest within the 3D data volume is additionally selected. An average for the plurality of vector dips is computed and a flat spot direction vector for each structural dip and a flat spot direction vector dip magnitude is determined. Sequences of distances from each reliability location is defined and interpolated onto the direction of the flat pot direction vector. The sequences of interpolated seismic attribute data is summed and the summed sequences of interpolated seismic attributed data values are stored.

Abstract: A system and method using inverse-vector processing to iterate through a loop of three steps: set a guide direction, invert opposite vectors, and average vectors to update the guide direction, for smoothing seismic amplitude data. The inverse-vector method can overcome instabilities where the traditional structure-tensor approach fails. The inverse-vector smoothing is simple to implement and more computational efficient. The resultant dips and azimuths are spatially consistent and thus more interpretable and suitable for calculation of curvature and other dip based attributes.

Abstract: Implementations described herein provide methods and devices for interpolating or estimating data from previously acquired data. Furthermore, implementations described herein calculate optimum interpolation operators by maximizing the spatial bandwidth of interpolation operators within a specified acceptable mean square error. According to one implementation, spatial bandwidth may be maximized by selecting a local grid within a global grid having nodes corresponding to desired interpolation locations. According to another implementation, spatial bandwidth may be maximized by specifying maximum wave numbers when calculating interpolation operators.

Abstract: Method for predicting lithology and porosity of subsurface rocks from seismic reflection data. The seismic data is inverted to yield elastic properties of the rocks such as the compressional and shear impedances. A rock physics model is built to relate porosity, the shale volume fraction, the fluid content of the rock and the elastic properties of the rock. The model is run backward in a second inversion process to solve for porosity and lithologic properties such as the shale volume fraction.

Abstract: The present invention provides methods and systems for microseismic hydraulic fracture monitoring in real-time. The methods and systems of the present invention may include continuous map migration of recorded microseismic signals. The methods and systems provide robust automated simultaneous detection and location of microseismic events.

Abstract: A method is disclosed for depth migrating seismic data. The method includes pre-stack time migrating the seismic data to form a stacked, time migrated image. The stacked, time migrated image is demigrated, and post-stack depth migration is then performed on the demigrated image. In some embodiments, the pre-stack time migration and the demigration account for ray bending in vertically transversely isotropic media.

Abstract: A method is provided for building a subsurface velocity model from a plurality of bins of marine acquired seismic traces of the subsurface.

Abstract: A method is provided for predicting a value of a designated rock or fluid property in a subterranean geologic volume. A first predicted value of the designated rock or fluid property is also assigned to a volume of a multi-dimensional, multi-scale model. A first predicted value of seismic response for the model volume is calculated from a response model using the first predicted value of the designated rock or fluid property, wherein the response model is responsive to changes in predicted values of the designated rock or fluid property. A synthetic trace is generated and iteratively compared to the corresponding trace obtained from one or more sets of actual seismic data to determine a difference while consistency is maintained between the types, scales and dimensions of values and data.

Abstract: Seismic data recorded by subsurface seismic sensors placed in a borehole, such as an oil or gas well, are transformed via a process of upward wavefield propagation to pseudo-receivers at the surface of the earth. The seismic data thus transformed can be processed as though the data had been recorded by the pseudo-receivers at the surface rather than in the borehole where the data were actually recorded. This method accurately accounts for seismic source statics, anisotropy, and all velocity effects between the real receivers in the borehole and the pseudo-receivers at the surface of the earth.

Abstract: System and method for performing anisotropy corrections of post-imaging seismic data for a subsurface formation. The method may receive seismic data, preferably pre-stack seismic data comprising a plurality of traces, e.g., collected from a plurality of source and receiver locations. The seismic data may be imaged/migrated to produce imaged seismic data, which is organized into an arrangement that preserves aspects of the relative seismic propagation angle in the subsurface. One or more anisotropic parameters and corresponding corrections may be determined by analyzing the organized imaged seismic data. The determined parameters or corrections may be used to correct at least a subset of the imaged seismic data, thereby producing corrected seismic data which is useable in analyzing the formation. The corrected data may then optionally be stacked to produce a collection of corrected stacked traces, and/or analyzed as desired. The pre-stack data and/or the corrected seismic data may optionally be displayed.

Abstract: A method for seismic event mapping includes transforming seismic signals recorded at selected positions into a domain of possible spatial positions of a source of seismic events. An origin in spatial position and time of at least one seismic event is determined from space and time distribution of at least one attribute of the transformed seismic data.

Abstract: The method of the invention allows, by means of an iterative inversion algorithm, prediction of the spatial distribution of the lithologic composition of sediments deposited in a sedimentary basin during a geologic time interval, and the temporal evolution of the depositional profile throughout filling of the basin, while respecting exactly the thicknesses of the sedimentary sequences measured otherwise which is useful for locating hydrocarbon reservoirs. Input data of thickness maps of the sedimentary layer, location and composition of the sediment supply at the boundaries of the sedimentary basin, and physical parameters characterizing transport of the sediments during the period of consideration are applied to an iterative inversion loop initialized by the accommodation provided by a stationary multilithologic diffusive model which works as a fixed-point algorithm correcting the accommodations by means of a preconditioning of the residue on the sequence thicknesses obtained by an instationary model.

Abstract: A system and method for perceiving drilling learning through visualization is provided. In one embodiment using three-dimensional visualization of the earth model as a foundation, a new IT development strategy focuses on perceiving “Drilling Learning” by an intuitive method. Symbols, known as “Knowledge Attachments” are attached to each wellbore trajectory displayed in the three-dimensional environment, with each symbol indicating a specific event—such as one related to drilling operations or problems. A Knowledge Attachment system proves particularly useful to represent disparate data at once, in such a manner that the interdependencies between the earth model and drilling operational data are evident and correlated. Operational issues and lessons learned from prior wells are easily accessed and perceived in the context of the earth model. By understanding this information at the beginning of the well planning process, operational efficiencies may be improved.

Abstract: The invention concerns a method for extracting a geological horizon and related properties of a seismic representation, comprising a step (100) which consists in digital modeling with continuous local seismic traces, calculating the optimal offset and defining a conditional neighbourhood of a reference central continuous local seismic trace; a step (101) which consists in defining a two-dimensional matrix whereof the line and column indices correspond to the coordinates of the geophones; a third step (102) which consists in selecting a seed point; a fourth step (103) which consists in determining the point vertically closest to the seed point and a fifth step (104) which consists in assigning to the point P(p,q,t) the value P(p,q,t+hij,pq,k), where hij,pq,k is optimal offset of the neighbouring point P(i,j,k), so as to estimate the related properties of the conditional neighbourhood thereby filling the two-dimensional extraction matrix of step (101).

Abstract: Method for building a subsalt velocity model. In one implementation, the method may include determining a velocity of a sedimentary area surrounding a salt body, determining an initial velocity of a subsalt sedimentary area disposed below the salt body by interpolating the velocity of the surrounding sedimentary area with one or more boundaries defining the salt body, and updating the initial velocity of the subsalt sedimentary area.

Abstract: A method having application notably to the development of petroleum reservoirs for fast generation of a geostatistical reservoir model on flexible grid, representative of a porous heterogeneous medium. A flexible grid best discretizing the medium is first generated. A regular Cartesian grid whose cell size, in a given direction, is greater than or equal to the largest size of the cell of the flexible grid in the same direction is then generated. On this Cartesian grid, at least one realization of petrophysical quantities characteristic of the heterogeneous medium is simulated from a geostatistical simulator. Then, this realization is resampled with a smaller discretization interval in order to know the petrophysical value associated with the realization at any point of the medium. Finally, the values of the simulated petrophysical properties are assigned to the center of the cells of the flexible grid.

Abstract: Implementations of various techniques for a seismic inversion method. In one implementation, the method may include estimating a wavelet of a recorded seismic trace at a well location, transforming the recorded seismic trace to a relative seismic impedance trace, estimating a low frequency model trace using the relative seismic impedance trace and adding the low frequency model trace to the relative seismic impedance trace to generate an absolute trace.

Abstract: The invention concerns a method for smoothing a subsurface property in a geological structure represented by seismic measurements comprising a step (100) which consists in digital modelling by continuous local seismic traces, calculating an optimal offset and defining a conditional neighbourhood of a reference central continuous local seismic trace; a step (101) which consists in selecting the property to be smoothed on a conditional neighbourhood; a third step (102) which consists in substituting properties of the conditional neighbourhood with offset properties; and a fourth step (103) which consists in selecting an average of the properties offset on the conditional neighbourhood at step (102).

Abstract: A system and method for interpreting reverse faults and multiple z-valued seismic horizons. A platform executable on a computer is provided for interpreting seismic data. The platform includes a user input module for accepting user inputs related to interpretation of seismic data and also includes a graphics processing module for displaying to a user graphics related to seismic data. A reverse fault module is provided for enforcing rules governing acceptable multi-z horizon picks. Such rules allow a multiple z-valued horizon to be interpreted as a single horizon. The invention further includes methods for allowing a user to interpret seismic data in a reverse-fault environment. An interpretation of seismic data is displayed and reverse faults and multiple z-valued horizons are created and displayed within the interpretation. A set of reverse-fault environment properties is associated with the reverse faults and the multiple z-valued horizons.

Abstract: Methods for scanning geophysical data sets to find geological entities with specific geophysical responses entail selecting a focus sub-volume and a background sub-volume proximal to the focus sub-volume. The sub-volumes include discrete sampling locations and are located within at least two geophysical data sets. Each discrete sampling location has associated data values. The background data volume and each data value are normalized. A determination is made to whether a data value is inside or outside of the background data. A distance value is associated with each determination. The distance values are evaluated to find discrete sampling locations with specific geophysical responses. The anomalous data points can be related to the presence of hydrocarbon or water bearing strata at the corresponding depth locations of the data points.

Abstract: A method for estimating and/or reducing uncertainty in reservoir models of potential petroleum reservoirs comprises receiving the results of a stochastic seismic inversion, and transforming the inversion data into a form suitable for reservoir modelling and flow simulations, while honoring inter-property and inter-layer correlations in the inversion data as well as measured well data and other geological constraints.

Abstract: A method for automated extraction of surface primitives from seismic data is presented. A preferred embodiment of the method includes defining, typically with sub-sample precision, positions of seismic horizons through an extrema representation of a 3D seismic input volume; deriving coefficients that represent the shape of the seismic waveform in the vicinity of the extrema positions; sorting the extrema positions into groups that have similar waveform shapes by applying classification techniques with the coefficients as input attributes using unsupervised or supervised classification based on an underlying statistical class model; and extracting surface primitives as surface segments that are both spatially continuous along the extrema of the seismic volume and continuous in class index in the classification volume.

Abstract: A method of investigating a reservoir region in a subsurface formation by a time-lapse seismic survey. The subsurface formation comprises a further formation region adjacent to the reservoir region. Data are obtained from a time-lapse seismic survey and includes seismic data of the subsurface formation at a first point in time and a later point in time. The seismic data is processed to obtain a seismic representation of change in a predetermined seismic parameter in the further formation region, whereby the seismic parameter is dependent on stress. The seismic representation of change in the seismic parameter in the further formation region is interpreted for an indication of changes of stress distribution in the further formation region, and a property of the reservoir region is derived using the indication of change of stress distribution in the further formation region.

Abstract: Method and apparatus for predicting surface multiples, which includes (a) selecting a target subsurface line (SSL); (b) selecting an input SSL within an aperture of the target SSL; (c) selecting a point on a line twice the distance between the input SSL and the target SSL, the point corresponding to a potential downward reflection point of the surface multiples for a trace; (d) generating a potential surface multiple for the trace corresponding to the point; (e) repeating steps (c) through (d) for each point on the line to generate an inline of potential surface multiples corresponding to each point on the line; (f) repeating steps (b) through (e) for each input SSL within the aperture of the target SSL to generate potential surface multiples for the trace corresponding to each input SSL within the aperture; and (g) adding the potential surface multiples to generate a surface multiple for the trace.

Abstract: A Method for Interpreting Seismic Data Using Duplex Waves is described. Whereas substantially horizontal boundaries are readily identified using conventional processing methods based on primary reflections, it has been difficult or impossible to use such methods for locating substantially vertical events or boundaries. The method of the present invention uses secondary reflections to locate substantially vertical events by gathering common source or receiver traces for processing. Wave fields of these gathers are continued downward to the level of the base boundary, then at each discrete depth level, a seismic image of sub-vertical events is formed. The downward-continued gathers correspond to the travel time of the wave from when it left the source or the receiver point, was reflected from the base boundary, and arrived to the corresponding point of the discrete level of wave-field continuation.

Abstract: A method of seismic data analysis to provide clustering of A.V.O. data into A.V.O. anomaly types, the method comprising: obtaining successive values of a plurality of seismic attributes, each seismic attribute comprising a respective property of a seismic reflection event, grouping said values using a running window of a predetermined size into a plurality of groups, for each group identifying first and second parameters corresponding to said first and second attributes, and plotting each group as a single event based on said group parameters, said group parameters having been selected to cause clustering of said seismic reflection events on said plot according to the presence or absence of A.V.O. anomalies.

Abstract: A method for multipath imaging of three-dimensional volumes of seismic data, given the subsurface velocity distribution. Ray tracing is used to generate the volumes of information needed for migration, but using coarse-scaled grids for computational efficiency. Interpolation methods are disclosed that address the specific shot points and receiver locations and that enable a final image on a grid with the desired resolution. Data storage and retrieval techniques are also disclosed.

Abstract: A method for seismic characterization of subsurface Earth formations includes determining at least one of compressional velocity and shear velocity, and determining reservoir parameters of subsurface Earth formations, at least including density, from data obtained from a wellbore penetrating the formations. A quality factor for the subsurface formations is calculated from the velocity, the density and the water saturation. A synthetic seismogram is calculated from the calculated quality factor and from the velocity and density. The synthetic seismogram is compared to a seismic survey made in the vicinity of the wellbore. At least one parameter is adjusted. The synthetic seismogram is recalculated using the adjusted parameter, and the adjusting, recalculating and comparing are repeated until a difference between the synthetic seismogram and the seismic survey falls below a selected threshold.

Abstract: Disclosed herein are systems and methods for calibrating seismic data using seismic attribute pairs such as amplitude versus offset (AVO) attributes. In one embodiment, a method includes: receiving traces from a seismic survey that covers a volume; determining a pair of seismic attributes for each of a plurality of positions within the volume (“survey pairs”); deriving pairs of seismic attributes from a source independent of the survey (“independent pairs”); calculating a rotation angle and one or more scale factors that calibrate the survey pairs relative to the independent pairs; and calibrating the traces with the rotation angle and the one or more scale factors. Thus, for example, the overall amplitude and relative AVO of pre-stack seismic gathers can be calibrated to match that of synthetic gathers by matching crossplots of AVO attributes generated from each dataset.

Abstract: An integral evaluation method for a complete numerical gauging of work pieces includes building an integral objective as a polynomial composed of two different objective groups and multiplying each group of the objectives with different free adjustable weighing parameters. The objective groups are Chebychev objectives group and a least-squares objectives group. The free adjustable weighing parameters include all prescribed connections between the elements with a tolerance requirements according to a drawing of the elements and a weighing of the elements.

Abstract: A method of obtaining a spatial model of a property of part of a subsurface formation located between underground seismic receivers in which at least two sets of pairs of seismic receivers are utilized and one pair of receivers is used to record a signal from a seimic source and obtaining a response by solving (s11(?t){circle around (x)}s21(t))=r11,21(t){circle around (x)}(s11(?t){circle around (x)}s21(t)), wherein the symbol {circle around (x)} denotes convolution and wherein s11(?t) is the time-reverse of the signal s11(t). A path-related attribute is selected from transmission response r11,21(t) that corresponds to the property of the subsurface formation and a tomographic reconstruction technique is applied to the path-related attribute to obtain the spatial model of the property of part of the subsurface formation.

Abstract: Representative embodiments provide for a computer including a program code configured to cause a processor to invert and thereafter calibrate first and second data sets, subtract the inverted second data set from the inverted first data set to derive a time-lapse data set, calculate a model including a plurality of parametric values, sort the plurality of parametric values into a plurality of bins, select, map and calibrate a plurality of optimal parametric values from the plurality of bins, and plot the plurality of calibrated optimal parametric values to represent at least one physical characteristic of a subterranean reservoir of hydrocarbons. The method includes deriving a time-lapse data set from a first seismic data set and a second seismic data set, deriving a model, sorting the plurality of values into bins, selecting, mapping and calibrating a plurality of optimal values from the bins, and plotting the calibrated values.

Abstract: The invention is a method for performing a stratigraphically-based seed detection in a 3-D seismic data volume. The method incorporates criteria that honor the layered nature of the subsurface so that the resulting seismic objects are stratigraphically reasonable. The method may be used to extract from a seismic data volume all seismic objects that satisfy the input criteria. Alternatively, the method may be used to determine the size and shape of a specific seismic object in a seismic data volume.

Abstract: A number of seismic stacks are precomputed (20) for known velocity fields. The velocity fields are chosen to span the range of velocities of interest. The stacks are then arranged (21) in the 3D memory of a graphics computer (10–14) using time and position as first dimensions and the index of the velocity field as the last dimension. In such 3D space, any velocity field to be used for stacking appears as a surface (S) within a volume. Projecting the seismic stacks onto that surface provides the seismic line stacked for the velocity field of interest.