Statistical approaches to interpretation of seismic reflection data

Abstract

Statistical approaches offer a potentially powerful tool for interpretation of the complex seismic reflection patterns typically recorded from the crystalline crust. These methods focus on characterization of the complexity, scaling characteristics and spatial variability of the wave field with the ultimate aim of recovering information on the geometric variation of the geology. In this paper we examine the potential and limitations of statistical approaches, focusing on the issues of quality of the statistical estimator, how noise and migration affect spatial statistics and the potential for both qualitative and quantitative approaches to statistics-based interpretation. The limited spatial bandwidth of typical seismic data presently inhibits complete quantitative recovery of the spatial properties of the geology. Migration, which should increase the spatial bandwidth, has to date, not proven particularly effective on complex wave fields. Mapping of relative variation in the spatial properties of the wave field reveals subtle differences that are directly associated with the geometric character of the geology. Such mapping allows identification and mapping of variations in the geologic fabric at scales significantly larger than the seismic wavelength and objective access to scaling information that has previously been unavailable.