Stratigraphic prospecting methods have made extensive use of seismic reflection data during recent years. Although many of these qualitative concepts have helped geoscientists define lithologic patterns and their characteristics with improved accuracy, quantification of the methodologies has been difficult. Multivariate statistical methods are useful in delineating and characterizing wave-form patterns from a multiplicity of seismic reflection lines. Recursive factor-analysis methods are used first to identify the number of wave-form patterns that exist in a particular zone of seismic data and then to assign each seismic trace to a particular wave-form group (or seismic facies).

This analysis yields two products: average and end-member wave forms for each of the various groups and a distribution map of the classified wave forms. Resulting average and end-member wave forms can be used in conjunction with well information and/or seismic models to infer a lithologic meaning for the seismic facies. The map of classified wave forms can effectively augment other geologic data and concepts in establishing environments of deposition and other distributional information. Further quantification of the wave-form patterns can be established through discriminant-analysis procedures. These resulting classification functions are useful when new data are integrated into the analysis and when correlating well information with the wave-form data.

End_of_Article - Last_Page 700------------