Incorporating complex geological features into geostatistical property modeling

Geostatistical modeling of reservoir properties is common in the assessment of expected reservoir flow. Typically, static properties such as porosity, saturation and permeability are modeled with geostatistical techniques to quantify the level of uncertainty in reservoir characterization. Lately, practitioners have become concerned that purely statistical techniques do not sufficiently integrate all geological knowledge available; rather statistical assumptions such as Gaussianity and simplistic assumptions on geometry such as global directions of anisotropy or simple stratigraphic coordinates dominate. The proposed methodology can incorporate complex information regarding deposit geometry in the form of locally varying anisotropy (LVA) to generate numerical property models that better reflect the level of geological knowledge available. Specifically, surfaces from seismic data, geological interpretations and dip meter data can be used to assess local orientations of continuity in the deposit and generate an appropriate LVA field. This work presents a novel technique for integrating faulting and LVA information in a single numerical modeling step.