One of the biggest roadblocks to making accurate 3D block models of numeric drilling variables is the difficulty of constraining the anisotropy (i.e., directional dependency) of a particular attribute. Due to the inherent structural nature of geological or mineralised systems, it is important to consider anisotropy when defining the preferential directional interpolation weighting for grade estimations (interpolation). Detailed anisotropy analysis can be achieved quickly in DRIVER using machine learning (ML).

This example drilling dataset is from a Sn project in Morocco. The main Sn ore zone is hosted in moderately NW-dipping inclined shear zone.

The Achmamch Sn deposit in Morocco, showing the concentration of Sn within moderately northwest-dipping shear bands (red box).

 

DRIVER can automatically determine the optimal anisotropy for each of the elements in the dataset simultaneously using ML. For this process, we open the ‘New global anisotropy estimation’ tool and then multi-select all of the different drilling assays. DRIVER will calculate default parameters-related to the 3D distribution of the dataset (e.g., average horizontal drill spacing). For this example we will only adjust the ‘minimum feature length’ parameter to force the anisotropy algorithm to evaluate for high-aspect ratio features (like shear zones). Clicking ‘Submit’ starts the processing of each data attribute on the cloud.

The ‘New global anisotropy’ tool in DRIVER

 

For this dataset, the anisotropy objects finished processing within several minutes. Each object can now be dragged and evaluated in the main 3D window. For the Sn, we can see the anisotropy (represented as an ellipsoid) following the strike and dip of the shear zones (255/30).

The calculated anisotropy ellipsoid for Sn, an SW-NE elongate ellipsoid, dipping moderately NW.
The stereonet for this ellipsoid is shown in the lower right corner.

 

The anisotropy of each attribute in DRIVER is evaluated independently, with the results presented to the user as a stereonet map showing a heatmap showing the relative strengths of the 3 principal axes. DRIVER also evaluates the anisotropy range parameter using a variogram oriented down the major principal axis.

3D block model of interpolated Sn grade using anisotropy configuration for Sn.
Note the NW-dipping lenses of high-grade Sn.

 

Aside from aiding a high-level, data-driven understanding of the deposit structure and geochemistry, the primary purpose of this detailed anisotropy analysis in DRIVER is to inform the 3D interpolations of attribute concentration into block models. This will be covered in the next DRIVER Primer article, Block Models.