Description

The Multisource RTM module performs Reverse Time Migration (RTM) using a multisource blending strategy on a single compute node. Multiple shot records are grouped into super-shots and back-propagated simultaneously through the depth velocity model using a two-way wave equation finite-difference scheme. The cross-correlation imaging condition is applied at each time step to produce the final depth-domain image. Random phase encoding reduces crosstalk noise between blended sources in the stacked image.

RTM is the most accurate depth migration algorithm because it correctly handles turning waves, prismatic events, and steeply dipping reflectors. Use this module when conventional Kirchhoff or one-way wave equation migration produces poor imaging of structurally complex targets such as thrust belts, salt flanks, or gas chimney flanks.

Input data

Input SEG-Y data handle

Link to the SEG-Y file handle providing the pre-stack shot records to be migrated. Data should be pre-processed (noise attenuated, multiples attenuated, statics applied) before input to RTM.

Input data trace headers

Link to the trace header vector containing source and receiver coordinates used to position the RTM operator correctly in the model space.

Input depth velocity

Link to the depth-domain interval velocity model used for propagating the forward source wavefield and back-propagating the receiver wavefield. The velocity model should be smooth at the scale of the dominant wavelength to suppress numerical dispersion.

Parameters

Datum

Reference datum elevation (m) to which the acquisition surface is referenced in the velocity model. Default: 0 m.

Time weight exponent

Exponent applied to a time-dependent gain function used to compensate for wavefield spreading in the source wave propagation. Default: 1. Valid range: 0–3.

Depth weight exponent

Exponent for a depth-domain gain applied to the migrated image to balance amplitudes between shallow and deep image zones. Default: 1. Valid range: 0–3.

Use random phase

When enabled, applies random phase shifts to each source before blending into a super-shot. This incoherently distributes crosstalk energy so that it averages toward zero when multiple super-shots are accumulated. Default: enabled. Disable only for testing with a single source.

Multisource group size

Number of shot records blended into each super-shot. Larger groups reduce computation time proportionally but increase crosstalk noise. Default: 1 (single-source mode, no blending).

Multisource group step

Index stride between sources selected into the same super-shot group. Spreading sources spatially reduces coherent crosstalk by distributing interference energy. Default: 1.

Multisource group distance

Minimum source-index distance between any two sources assigned to the same super-shot group. This enforces a minimum spatial separation between blended sources. Default: 1.

Wavelet params

Container group defining the source wavelet parameters used in the forward propagation step of RTM.

Frequency

Dominant frequency (Hz) of the Ricker source wavelet used for forward modelling. Default: 15 Hz. Set this to the dominant frequency of your field data for the most geophysically meaningful image.

Modelling params

Container group with parameters controlling the spatial and temporal grid of the finite-difference propagation engine.

Model resolution

Spatial grid cell size (m) for the finite-difference wave propagation. Default: 5 m. Must be smaller than one-fifth of the minimum wavelength in the model. Finer grids produce higher accuracy but require substantially more memory and time.

Number of depth samples

Number of depth grid cells in the modelling volume. Default: 1000. Set this so that the modelling depth (Number of depth samples × Model resolution) covers the full depth range of interest.

Boundary extrapolation range

Width (m) of the absorbing boundary layer surrounding the modelling grid. Default: 400 m. Absorbs outgoing energy to prevent edge reflections from contaminating the image.

Free surface

When enabled, a free-surface reflective condition is applied at the top boundary, including surface-related multiples in the modelled wavefield. Disable for primary-only imaging. Default: disabled.

Modelling step

Internal time increment (s) for each finite-difference propagation step. Default: 0.001 s. Must satisfy the CFL stability condition. Reduce this value if numerical instabilities appear in the output.

Number of modelling steps

Total number of propagation time steps. Default: 4001. Total simulation time equals Number of modelling steps × Modelling step. Set this large enough for wavefields to traverse the full depth of the model twice (down and back up).

P param

Regularisation parameter for the imaging condition normalisation. A small positive value (default: 0.05) stabilises the deconvolution-based imaging condition in areas of low source illumination. Increase this value to reduce amplitude anomalies near poorly illuminated zones; decrease it for more accurate amplitude preservation in well-illuminated areas. Valid range: 0–1.

References

Baysal, E., Kosloff, D. D. and Sherwood, J. W. C., 1983, Reverse time migration: Geophysics, 48, 1514-1524.

Romero, L. A., Ghiglia, D. C., Ober, C. C. and Morton, S. A., 2000, Phase encoding of shot records in prestack migration: Geophysics, 65, 426-436.