Global Exploration and Geotechnology, 1996 AAPG Annual Meeting, San Diego, California, May 19-22, 1996 (Poster Session)
AbstractWhen evaluating a site, whether for oil, minerals, or contaminants in ground water, a principle concern is the distribution of material properties. A traditional approach has been to apply geostatistical methods such as kriging or conditional simulation. These approaches are based on the assumption of stationarity (i.e. that the spatial variation of properties is consistent across the site). At many sites, the stationarity assumption is not valid and can lead to inaccurate results. One approach to circumvent this limitation is to divide the area into zones where the stationarity assumptions are reasonable, krige each zone, and manually merge the results together. This approach has three major draw backs, 1) boundaries between zones are abrupt, 2) the merging process is tedious, and 3) there is no way to manage 'gradational' boundaries.
An integrated system which allows a modeler to: 1) define multiple, distinct zones within a model; 2) define zonal inter-relationships (e.g. Zone A grades into zone B. Zone C and Zone D have a sharp contact), and model the results using simple or ordinary kriging, or conditional simulation is presented. This technique is integrated into a modeling package which allows users to examine basic site statistics, develop and model semivariograms, krige and simulate material properties, model ground water flow and contaminant transport, assess risk or uncertainty, and visualize results with 2D, 2-1/2D, and 3D tools.