Depth Conversion | Oil Authority

Depth conversion is the process of transforming seismic reflection data from the time domain (in which it is acquired, with seismic two-way travel time in milliseconds as the vertical axis) to the depth domain (with formation depth in meters or feet as the vertical axis), enabling structural maps that reflect the true geometric positions of formation boundaries in the subsurface rather than the time-distorted geometry that would be interpreted directly from time-domain seismic; the conversion is performed by applying an interval velocity model that specifies how fast seismic waves travel through each formation interval, integrating the velocity model from surface to each mapped horizon to calculate the cumulative two-way travel time corresponding to each depth, and using that time-depth relationship to rescale the seismic in depth; depth conversion accuracy is critical because time-domain seismic structural maps are distorted by lateral velocity variations (the pull-up effect beneath fast formations and pull-down beneath slow formations) that can create false or exaggerated structural closures that appear to be commercial prospects on time maps but flatten out or disappear on accurately depth-converted maps, making depth conversion accuracy directly relevant to exploration success rates and the economic value of seismic data.

Key Takeaways

Velocity model construction for depth conversion integrates three primary data sources with different spatial resolution and lateral coverage: seismic velocity analysis (root-mean-square velocities derived from normal moveout analysis across the 3D seismic volume that provide laterally continuous but low-vertical-resolution interval velocity estimates throughout the seismic survey area), well-log velocity data (sonic logs and check-shot surveys that provide high vertical resolution velocity at the specific well location but no lateral coverage away from the well), and regional velocity trends (empirical velocity-depth relationships derived from regional well control and calibrated to the specific geological setting); interval velocities calculated from seismic RMS velocities using the Dix equation (Vi = sqrt((V_rms2^2 × t2 - V_rms1^2 × t1) / (t2 - t1))) provide the initial velocity model for depth conversion, which is then calibrated to the well-log velocities by applying a low-frequency correction that adjusts the seismic-derived velocities to match the sonic log measurements at each well, interpolating the correction laterally across the 3D volume using kriging, cokriging, or other spatial interpolation methods.
Check-shot surveys provide the most direct velocity calibration for depth conversion — a check-shot is run in a well by positioning a seismic source at surface and recording the one-way travel time of the seismic wave to a downhole geophone tool at multiple depths, directly measuring the average velocity from surface to each check-shot depth; the check-shot one-way time measurements combined with the known well depth at each measurement level define an exact time-depth relationship at the well location that is the primary calibration constraint for the velocity model; the difference between the check-shot derived time-depth relationship and the time-depth relationship predicted by the initial seismic-velocity-derived model at the well location is the "depth residual" — the local velocity model error that must be corrected to make the depth conversion honor the well data; in 3D seismic surveys with multiple wells, the depth residuals at all wells are used simultaneously to calibrate the velocity model, with residual values at well locations constrained exactly and the corrections interpolated or spatially estimated between wells.
Vertical seismic profile (VSP) data provide both check-shot calibration data and upgoing reflection data that are essential for high-quality depth conversion — a zero-offset VSP (with the source directly above the well) provides the same time-depth calibration as a check-shot but also records the complete upgoing reflection wavefield that allows identification of specific seismic reflectors from the well perspective, directly correlating log-defined formation tops to seismic events at precisely known depths and two-way times; walkaway VSP and three-dimensional VSP geometries provide additionally the near-well seismic velocity field including anisotropy (directional velocity variations common in shale formations), allowing the velocity model to be corrected for the effects of anisotropy near the wellbore that cause seismic moveout velocities to differ from the true vertical velocities needed for depth conversion; VSP data are particularly valuable in areas with complex near-surface velocity variations (permafrost, shallow gas, rugose sea floor) where the near-surface velocity structure significantly distorts the seismic time domain and the VSP provides the direct subsurface velocity measurement that can constrain near-surface model corrections that cannot be reliably derived from standard seismic velocity analysis.
Pull-up and pull-down effects on time-domain seismic arise from lateral velocity contrasts that cause the seismic two-way time to a deep reflector to vary laterally even when the reflector's true depth is flat — a high-velocity formation (carbonate, salt, igneous intrusion) in the shallow section transmits seismic waves faster than surrounding low-velocity shales, causing the seismic waves to arrive earlier beneath the fast body and making the underlying reflectors appear higher (at earlier time) than they actually are (the pull-up effect); a low-velocity body (low-pressure shallow gas, low-velocity coal seam, overpressured shale) causes the opposite pull-down where underlying reflectors appear deeper in time than they actually are; depth conversion corrects these distortions by applying the correct interval velocities through each formation, eliminating the false structural relief that time-domain maps might interpret as closure — the Ekofisk chalk field in the North Sea is a classic example where the high-velocity chalk creates strong pull-up of the underlying reflectors, and depth conversion is essential to correctly evaluate the structure of formations beneath the chalk.
Iterative depth conversion using geostatistical techniques (kriging, sequential simulation, co-kriging with seismic velocity attributes) propagates velocity uncertainty through to depth uncertainty in a way that allows quantification of the structural uncertainty at each mapped horizon — the resulting depth uncertainty maps identify which areas of the structure have well-constrained depths (near wells where the velocity model is calibrated) and which areas have large depth uncertainty (far from wells where only seismic velocity analysis constrains the model); for exploration prospects, the depth uncertainty maps directly affect the estimated prospect volume by showing how the structural closure area and relief vary across the uncertainty range of the depth conversion, with larger depth uncertainty producing larger ranges of possible closure areas and hydrocarbon column heights; the prospect volume probability distribution derived from depth uncertainty analysis is a required input to SEC reserve reporting methodology for prospects under evaluation and to risking frameworks used by major operators to prioritize exploration portfolios.

Fast Facts

The development of 3D seismic technology in the 1980s and 1990s dramatically increased the importance of accurate depth conversion by providing laterally continuous subsurface images that could be depth-converted volumetrically rather than only along isolated 2D seismic lines. The first 3D marine seismic surveys in the North Sea and Gulf of Mexico in the early 1980s demonstrated that 3D depth-converted structural maps could locate reservoir tops with accuracy of ±20 to 50 meters in moderate-velocity-complexity areas, sufficient to commit to development well locations without additional appraisal drilling. Today, in areas of high velocity complexity (subsalt deepwater GoM, pre-salt Brazil), depth conversion accuracy has become the primary technical challenge in exploration, with depth residuals at wells of ±100 to 300 meters indicating that velocity model improvement — through better tomographic inversion, waveform inversion, and geological model constraints — remains the frontier problem in applied seismic geophysics.

What Is Depth Conversion?

Seismic reflection data records the time it takes for sound waves to travel down through the earth, reflect off rock boundaries, and return to surface hydrophones or geophones. The result is a seismic section with formation boundaries displayed at their two-way travel time rather than their actual depth. In a simple, uniform-velocity earth, converting from time to depth would require multiplying time by a single velocity constant. In the real earth, velocity varies dramatically with formation type, depth, pressure, and mineralogy — limestone travels at 6,000 m/s while shale travels at 3,000 m/s — and these variations cause time-domain seismic images to be geometrically distorted in ways that can create apparent structural highs where none exist.

Depth conversion applies a velocity model to undo this distortion. The resulting depth map shows formation boundaries at their true positions in the subsurface. For an exploration prospect, this is the difference between a structure with 100 meters of closure (perhaps economic) and one with 20 meters of closure (probably not) — the time map may show the larger closure while the depth map reveals the smaller one, after correcting for the velocity pull-up caused by a fast formation above the prospect. Getting depth conversion right is not a cosmetic improvement on the seismic — it is the transformation that makes the seismic a reliable guide to where the reservoir actually is.

Depth Conversion Methods and Uncertainty

Map migration and depth conversion sequence in 3D seismic processing has evolved through three generations: time migration + depth conversion (the traditional sequence where time-migrated seismic is interpreted and the resulting time-domain structural maps are depth-converted using a velocity model), depth migration without model update (depth Kirchhoff or wavefield-extrapolation migration using an initial velocity model that is not iteratively updated), and full-waveform inversion (FWI) with iterative velocity model update (the modern approach where the velocity model is iteratively refined by minimizing the difference between synthetic seismograms computed from the model and the actual recorded seismic waveforms, producing the most accurate velocity model and therefore the most accurate depth conversion); FWI has transformed depth conversion accuracy in subsalt and sub-basalt areas by recovering the long-wavelength velocity structure beneath complex velocity boundaries that earlier tomographic methods could not image, reducing depth prediction errors from ±200 to 300 meters to ±50 to 100 meters in the deepwater GoM subsalt environment.

Acoustic impedance inversion and rock physics calibration provide additional constraints on the velocity model for depth conversion in areas where the reflection coefficient at formation boundaries is dominated by density contrasts rather than velocity contrasts — log-derived acoustic impedances (product of velocity and density from sonic and density logs) calibrated against seismic amplitudes allow the structural interpretation and velocity model to be jointly constrained by both the amplitude information (which reflects impedance contrasts at boundaries) and the travel time information (which reflects integrated interval velocity through the section); this joint constraint approach reduces non-uniqueness in the velocity model inversion problem and typically produces more accurate depth predictions at wells not used in the calibration, providing independent validation of the depth conversion methodology before committing development well locations based on the depth-converted structure.

Depth Conversion Across International Jurisdictions

Canada (AER / WCSB): WCSB seismic depth conversion for conventional oil and gas exploration in Alberta and Saskatchewan typically involves relatively simple velocity models because the shallow stratigraphy is dominated by laterally continuous clastic sequences (Cretaceous Colorado Group, Mannville Group) with moderate velocity variations that do not create severe depth conversion uncertainty; AER requires that exploration well locations be approved based on a demonstrated geological target and structure, with 3D seismic depth conversion forming the primary structural evidence for the wellbore location's ability to test the proposed prospect; deepwater equivalents in WCSB exploration are absent, but the mountainous foothills belt in southwestern Alberta and northeastern BC presents severe depth conversion challenges due to complex thrust belt structure, high-velocity carbonates, and erratic near-surface velocity variations from glacial deposits that require sophisticated velocity model building techniques similar to those used in deepwater environments.