Scalable Earth Modelling: A Grid-less Multi-scale Multi-resolution Modelling Approach

For the last forty year, the same workflow has been used for generating static geomodels. First, a three-dimensional grid is built from a structural framework. Then well log and seismic data are upscaled to the grid cell resolution, and properties are modelled using geostatistical techniques that require inferring parameters (e.g., the variogram model) from these upscaled data. The grid is built at a user-specified scale and resolution; modifying the grid scale to focus on a different area of interest, or the grid resolution to capture additional geological detail, requires starting the geostatistical modelling workflow again from scratch. As a result, static geomodels are built for a unique purpose, creating silos between disciplines, and leading to inconsistencies among geomodels. Also, locking the grid resolution early in the modelling process results in the conservative choice of fine resolution grids to incorporate detailed geological features in case they would impact project forecasts, making the modelling process unnecessarily complex and time consuming.

A new grid-less modelling approach, named Scalable Earth Modeling, is presented. It enables fully integrated and consistent multi-scale modelling while allowing geomodelers to add geological detail only when needed, based on the accuracy of the forecasts required for making sound project decisions.