Velocity modeling

Velocity modeling software

Turn seismic velocities and well control into a depth framework you can drill against. g-Space builds the depth velocity model through one wizard with eight methods — from a single constant velocity to checkshots, markers and horizons solved together — then calibrates it to the wells and reports how far off it still is.

Download free trial The eight methods Watch Demos

Eight methods

One Create velocity model wizard

Checkshots & markers

Kriged onto the horizon framework

Pine chart

Model against well data, by depth

Depth velocity model in g-Space displayed along an arbitrary line through four wells, beside a location map showing the well positions and the line path

The Create velocity model wizard in g-Space listing its building methods, with a diagram of interval velocities derived from the time-depth relationship between horizons

In view

Depth velocity along an arbitrary line

A velocity model is how a project stops being two projects. The seismic knows time, the wells know depth, and the velocity model is the statement that reconciles them — it describes how velocity varies with position and depth across the survey, and g-Space uses it to convert interpreted data between the time and depth domains, to build depth maps and cubes, and to support depth imaging and structural interpretation. Which method you pick is a question about your data, not about the software: sparse wells and good seismic velocities take one route, a mature field with checkshots and markers in every well takes another, and the same wizard covers both ends.

Building methods

Nine ways to build it, ordered by what you have.

The Create velocity model wizard runs from the simplest method to the most data-intensive. Every one of them writes the same product — interval and average velocity cubes with depth and time maps — so the choice is only about which constraints you can actually bring. A ninth route skips the wizard altogether and brings in a model built elsewhere.

No well control

Constant

One interval velocity applied to the whole cube. A first time-to-depth relationship for feasibility work, and the sanity reference every more elaborate model gets compared against. Not for production depth conversion — it flattens any real depth trend.

Structure known

Constant layered

Piecewise-constant layers, each bounded by a horizon, a time map, a constant time or topography, each carrying its own interval velocity. Built for velocity provinces separated by mapped surfaces — salt over sediment, basalt over sediment.

Markers

Markers + horizons

Interval velocities derived from well-marker depths tied to time horizons, then kriged or triangulated between the wells. Velocity tracks the structure the horizons define while honouring the depth picks at each well.

Bin-aligned

Grid velocity by markers + horizons

The same marker-to-horizon ties, written on a bin grid rather than the horizons' native trace grid — for when 3D-grid creation or reservoir modelling downstream needs a regular, bin-aligned cube.

Checkshots

Checkshot extrapolation

Checkshot surveys turned into a full 3D interval velocity cube, layer by layer. Each layer gets its bounding surfaces, its internal layering type and either a constant value or a checkshot-derived curve, and the result is interpolated between wells and extrapolated beyond them.

Checkshots + markers

Checkshot + markers

Absolute depth-time calibration from the checkshots, blended with intermediate depth picks from the markers through a kriging interpolator. Markers give lateral control where checkshot sampling is too sparse to reach.

All three

Checkshot + markers + horizons

The most comprehensive method: checkshots for absolute calibration, markers for lateral variation and time horizons for the structural framework, all fed to one kriging solver. The production route in a mature field, and the one that reports tie quality per well.

Seismic velocities

Velocity volume

Builds the depth interval model from a velocity volume already in the project — 2D or 3D, time or depth, holding RMS, interval or average velocities. Declare which type it holds and g-Space converts accordingly, Dix-converting RMS to interval before resampling to depth.

Already built

Import velocity model

A model produced elsewhere loads straight from a SEG-Y file into the project's velocity model set, where it behaves like any other model — set as default, displayed, QC'd and exported back to SEG-Y.

What you control

The parameters that decide whether it is believable.

Between the wells is where a velocity model is invented, so the wizard puts the interpolation, the trend, the layering and the bounds in your hands rather than behind a default.

Interpolation you choose

Well-tie interpolation runs as Voronoi, ABOS, triangulation or kriging, with kriging the default. Kriging exposes its covariance model — spherical, Gaussian or exponential — and the correlation range that decides how far a well's influence reaches.

Trends that guide the fill

Each marker-to-horizon tie can take a trend to steer it laterally: none, an RMS velocity field, layer thickness or a data map, optionally inverted. A trend smoothing radius controls how gently that guide is applied.

Layering inside each interval

A layer is subdivided proportionally, parallel to its top or parallel to its bottom — the difference between a layer that thickens with the structure and one that hangs off a single surface.

Bounds and fallbacks

Minimum and maximum velocities clamp the cube to physically plausible values; a bottom velocity fills below the deepest tie and a constant velocity covers grid areas that neither checkshots nor markers constrain.

Smoothing where it earns its place

A high-cut frequency filter takes the noise out of a checkshot velocity curve before it is extrapolated, and post-interpolation smoothing works over trace and sample radii you set. A converted velocity volume gets its own vertical window and a lateral window that follows the project's time horizons.

Extent set deliberately

Datum, maximum depth and depth step shape the output cube, and the reference geometry can follow the intersection of the input maps, a polygon or a 3D survey with its own XY step or inline and crossline decimation.

Calibrated to the wells

Update depth velocity by well tie calibrates the model to the wells, adjusting velocities so the modelled depths match the well markers. Clear all well-tied maps removes those corrections again and restores the uncalibrated model.

Mismatch reported per well

Checkshot and marker methods write per-well logs of the difference between the model and the data that built it, and the combined method can compute them for every well in the project, not only the ones used in the calibration.

Workflow

Six steps from picked horizons to a depth framework.

Take stock of the constraints

Which wells carry checkshots, which carry the markers you would tie, and how far the time horizons reach. That inventory picks the method — the wizard is ordered from the least data-hungry to the most.

Open the wizard

Create depth velocity on the Velocity model bar opens the wizard; each method shows a preview image of what it does alongside its own parameter set and its layer or tie table.

Set the frame, then the ties

Datum, maximum depth, depth step and the velocity bounds first; then the layer table or the marker-to-horizon tie table, one row per layer or per tie, with a trend where the fill needs guiding.

Run it, or save it as a task

Run builds the model under a name you supply. Save to workflow instead keeps the entire setup — method and every parameter — as a Create velocity model task that can be reopened, edited and re-run.

QC it against the wells

Read the per-well mismatch logs, build a Pine chart from the model, and convert a time horizon to a depth map under each candidate model to see the two side by side in 3D.

Calibrate and put it to work

Tie the model to the wells, set it as the project default, and let depth conversion, depth maps, structural modelling and the depth 3D views all work from the same velocities.

QC & outputs

A model you can argue with, not just look at.

Every method writes the same shape of result, and every result can be displayed on the data that produced it — on time sections as time maps, on depth sections as surfaces, and in the 3D view as the model's own folder.

Checking the model

What g-Space gives you to judge it by

Pine chart

Plots the velocity model against well data with offset on the horizontal axis and depth on the vertical, so model and wells can be compared over the interval that matters. Built from the Velocity model bar.

Per-well mismatch

Checkshot and marker methods write logs of the difference between the model and the input data at each well; the combined method can extend that to every well in the project.

Model against model

Convert the same time horizon to a depth map under two different active models and compare the two maps in the 3D view — the difference is what the choice of method actually cost you.

On depth sections

The active model's depth surfaces draw on 2D lines, inlines, crosslines and arbitrary lines as a projection object, all layers at once or one map at a time.

On time sections

The model's time maps draw on time sections as lines at the times they hold along the section. Where a map has no values, its line is interrupted rather than drawn across the gap.

In 3D

The surfaces of the active model are collected in a Velocity model folder in both the time and the depth 3D view, alongside the rest of the project's objects.

What a run leaves behind

Stored in the Data Manager, ready for everything downstream

Velocity cubes

Interval velocity and average velocity, sampled on the trace grid, on a bin grid or on the geometry of the volume the model was converted from.

Depth & time maps

A depth and a time map for each layer of the model, written into the Maps folder where the rest of the project's grids live.

Marker horizons

The combined method can create interpreted horizons at the marker depths and include them in the output, so the well picks become mappable surfaces.

Time-to-depth conversion

The model is what converts interpreted data between domains, builds depth maps and cubes, and supports depth imaging and structural interpretation.

Model management

Several models live side by side under Velocity model; one is set as the default that everything downstream uses, and each can be renamed, removed or organised into folders.

SEG-Y in and out

An existing model imports from SEG-Y, and any model in the project exports back to SEG-Y for use outside g-Space.

Tools in this group

The parts of g-Space this topic is built from

Velocity model bar Create velocity model Velocity model QC Pine chart Update depth velocity by well tie Import velocity model Export to SEG-Y Maps Checkshots Markers Workflows

More g-Space capabilities

Every depth answer in the project runs through this model.

Depth-converted horizons, structural frameworks, 3D grids and volumetrics all inherit the velocities set here. These are the other topics in the g-Space workflow.

g-Space overview Geological interpretation Horizon picking Fault picking Well tie Seismic mis-tie Petrophysics Attribute analysis & AVO Geological modeling Reservoir characterization

FAQ

Questions, answered.

How does g-Space build a depth velocity model?

Through the Create velocity model wizard, opened from the Velocity model bar. You choose one of eight building methods, fill in its parameters and its layer or tie tables, then Run to build the model or Save to workflow to keep the whole setup as a task that can be replayed later. The methods run from a single constant interval velocity, through constant layers bounded by horizons, marker-to-horizon ties and checkshot extrapolation, up to a combined checkshot, marker and horizon solution. The finished model joins the project's depth velocity model set and becomes available for time-to-depth conversion and QC.

Can a velocity model be built from seismic velocities alone?

Yes — the Velocity volume method builds a depth interval model from a velocity volume already in the project. The volume can be a 2D line or a 3D cube, in time or in depth, holding RMS, interval or average velocities; you declare which type it contains and g-Space converts accordingly, Dix-converting RMS velocities to interval and resampling to depth. Vertical smoothing and lateral smoothing along the project's time horizons are available. It suits areas with good seismic velocity coverage and limited well data, and makes a sound starting cube to refine afterwards with a well-based method.

How is a velocity model checked against the wells?

Three ways. The Pine chart plots the velocity model against well data with offset on the horizontal axis and depth on the vertical, so model and wells can be compared directly. The checkshot- and marker-based methods write per-well mismatch logs between the model and the input data, and the fullest method can compute those for every well rather than only the ones used in the calibration. And because time horizons can be converted to depth maps using whichever model is currently active, two models can be compared by converting the same horizon under each and looking at the two depth maps in the 3D view.

What does a velocity model run produce?

Interval velocity and average velocity cubes, plus depth and time maps for the model's layers, all stored in the Maps folder of the Data Manager. Marker-based runs can additionally create horizons at the marker depths and write per-well time-depth mismatch logs. The model's depth surfaces display on depth sections and its time maps on time sections, and both appear in the 3D view. Models can be set as the project default, imported from SEG-Y and exported back to SEG-Y.

Get started

Build the depth model on your own wells.

Load your checkshots, markers and horizons, run two methods against each other and read the mismatch — take g-Space for a trial run, or talk to Geomage about a demo on your data.

Download free trial See pricing