Three-dimensional VSP elastic Kirchhoff pre-stack depth migration: Vinton Dome, Louisiana.

摘要

Vertical Seismic Profile (VSP) experiments are conducted to provide improved vertical resolution over conventional surface seismic data. While, in general, surface seismic data provides better horizontal resolution than VSPs, in certain geologic environments, VSPs can in theory provide improved resolution of steeply dipping reflectors that are poorly illuminated from the surface, such as salt flanks. The illumination of such a salt flank at Vinton Dome, LA, was the goal of a 3-D simultaneous surface/VSP acquisition program carried out by OPEX in 1999. Unfortunately, the service company used by OPEX was unable to provide a clear salt flank image.; The goal of my thesis is to use the Vinton Dome data, including 3-D interpretation of the seismic and well control, to better understand why the original VSP salt-flank imaging effort failed, and then to develop a new workflow and software to improved it.; Constructing a velocity model based on the surface seismic data, my wave equation models showed the salt dome flank is best illuminated by the horizontal components of the 3-D VSP, rather than by the vertical component imaged by the service company. These models also showed that VSP depth images from the sparse shot coverage suffered from considerable spatial aliasing resulting in images that look more like cabbages than geology. By using a newly developed semblance-weighted filter, I am able to significantly reduce these artifacts, and for the synthetic control data, produce relatively noise-free, though spatially limited images. The 2-D and 3-D modeling results indicated that the salt flank reflection is composed of both up-going and down-going events, such that the conventional VSP workflow of separating upgoing from down-going wavefields is inappropriate.; Finally, I developed an elastic Kirchhoff pre-stack depth migration algorithm that applies PP and PS polarization filters based on the emergence angle of each and every candidate reflector onto the VSP recording aperture. The results of these new developments applied to synthetic 2-D elastic and 3-D scalar wave equation models are excellent. The results applied to the real 3-D 3-C Vinton Dome data are encouraging, but less remarkable. While I had an exact velocity model for the synthetics, my field data velocity model was constructed from well control and surface seismic reflector picks designed to focus horizontal reflectors. I believe that an improved image will require a more accurate, tomographically-driven velocity model designed to focus the vertical salt flank.