Generating LVA Fields for use in Nonstationary Geostatistics

Recently there has been interest in incorporating locally varying anisotropy (LVA) into geostatistical models. The traditional assumption is that a deposit is stationary, that is, that the mean and variogram are constant in the modeling domain. Many natural deposits display non-stationary features and the assumption of stationarity is questionable. This paper does not discuss how the LVA field is used to build geostatistical models, the interested reader is revered to (Boisvert 2008). To implement the methodology in Boisvert (2008) it is necessary to parameterize the LVA field by the magnitude (anisotropy ratios) and orientation (major direction of continuity); an exhaustive LVA field is required for the deposit. This paper presents 6 different types of data that can be used to infer the LVA field as well as 5 methodologies to determine the LVA field from the data types (magnitude and direction). This paper provides practical methodologies to generate the LVA field from potential sources of data. Real data sources are used in this paper to highlight the proposed LVA field generation methodologies.