Footprint: A look at seismic acquisition geometries using 3D prestack depth migration

Gary F. Margrave

The acquisition footprint of a 3D survey refers to any imprint or modulation of the amplitude and phase of the final migrated depth image that is directly attributable to the survey geometry, that is, the positions of sources and receivers. Since processing algorithms differ in their ability to deal with geometric problems, the footprint is generally also a function of the processing flow. This paper examines some of the factors affecting footprint for a typical land 3D geometry when imaged with prestack source-record migration. As a preliminary, a brief overview is given of some of the sampling issues concerning possible spatial aliasing with respect to the sampling intervals of receivers, receiver lines, sources, and source lines. The footprint is then examined through a numerical simulation that consists of (1) image-source modeling of the reflection response of a uniform horizontal reflector, (2) prestack depth migration of each source record, and (3) summation of the migrated source records with and without illumination compensation. Two different models of illumination estimation are examined. One is the direct thresholding of the rectified migration response for each source record and the other is the thresholding of the rectified, normalized crosscorrelation of the source and receiver functions. Illumination compensation was undertaken by dividing the source-record stack by the corresponding stack of illumination estimates. Simulation results for both PP and PS recording are shown. Some general conclusions are: (1) some form of footprint is unavoidable, (2) direct stacking of migrated source record without illumination compensation leaves a strong aperture imprint (3) illumination compensation lessens the aperture imprint but can worsen the imprint of the geometry (4) the spatial aliasing due to course source and receiver line spacings is roughly compensated if the lines are orthogonal (5) the geometry is very effective against both random and coherent noise although the latter is more problematic (6) strong levels of coherent noise dramatically worsen the footprint effect. The Matlab script files for this simulation are released with this paper and are intended to allow exploration of parameter interactions that are beyond the scope of this introductory paper.