Methods to improve computer-assisted seismic interpretation using seismic attributes: Multiattribute display, spectral data reduction, and attributes to quantify structural deformation and velocity anisotropy

Computer-assisted seismic interpretation gained widespread acceptance in the mid 1980s that no 3D survey and few 2D surveys are interpreted without the aid of an interpretation workstation. Geoscientists routinely quantify features of geologic interest and enhance their interpretation through the use of seismic attributes. Typically these attributes are examined sequentially, or within different interpretation windows. In this dissertation, I present two novel means of presenting the information content of multiple attributes by a single image.; In the first approach, I show how two, three, or four attributes can be displayed by an appropriate use of color. I use a colorstack model of Red, Green, and Blue (RGB) to map attributes of similar type such as volumes of near-, mid-, and far-angle amplitude or low-, moderate-, high-frequency spectral components. I use an HLS model to display a theme attribute modulated by another secondary attribute, such as dip magnitude modulating dip azimuth, or amplitude of the peak spectral frequency modulating the phase measured at the peak frequency. Transparency/opacity provides a 4th color dimension and provides additional attribute modulation capabilities.; In the second approach I use principal component analysis to reduce the multiplicity of redundant data into a smaller, more manageable number of components. The importance of each principal component is proportional to its corresponding eigenvalue. By mapping the three largest principal components against red, green, and blue, we can represent more than 80% of the original information with a single colored image.; I then use these tools to help quantify and correlate structural deformation with velocity anisotropy. I develop an innovative algorithm that automatically counts the azimuth distribution of the fast P-wave velocity (or alternatively, the strike of the structural lineaments) weighted by the amount of anisotropy (or the intensity of the lineaments) at any point in the volume for a given analysis window. The output is a series of rose diagrams, which can be related to the variation of fracture intensity and azimuth across space and depth (or geologic time). I demonstrate that such automatically generated rose diagrams from an uninterpreted seismic volume can closely approximate those picked by an experienced interpreter. These lineaments are easily broken into fracture sets, each of which has its own azimuth, which I then correlate with water and oil production in the Kansas data volume. The results indicate that some fracture sets were open, karstified, and filled with shale in Mississippian time, resulting in closed fractures. A second fracture set was not altered during Mississippian time, and today are open fractures connected to the deeper water-bearing formations.