Common sampling biases
Diamond drilling is widely regarded as the “gold standard” for mineral exploration due to its ability to provide continuous, high-fidelity physical samples of the subsurface. However, even this precise method is susceptible to sampling biases that can significantly distort geological models and resource estimations. These biases generally stem from the interaction between the drilling equipment and the rock mass, or from the geometric relationship between the borehole and geological structures.
One of the most critical biases occurs during the extraction process. In heterogeneous orebodies, softer minerals (often the high-grade components like gold-bearing sulfides or carbonates) may be washed away by drilling fluids or ground down, while harder waste rock is preserved. This “preferential loss” leads to an underestimation of the grade if the mineralized portions are softer than the host rock (Dominy et al., 2019). Conversely, if the host rock is softer and preferentially lost, the resulting sample will show an artificial “grade enrichment.”
Boreholes are one-dimensional samples of three-dimensional space. A “blind zone” exists for any drilling program where structures (faults, veins, or joints) oriented parallel or sub-parallel to the drill hole are statistically under-represented (Fowler, 2013). This geometric bias is often referred to as the Terzaghi effect, where the probability of intersecting a feature is proportional to the sine of the angle between the borehole and the feature.
Bias can also be introduced during the “splitting” of the core. If the core is not cut exactly in half, or if one half is consistently chosen based on visual “richness” (a common human error), the resulting assay will not represent the true average of the interval (Dominy et al., 2019).
Correction and mitigation strategies
The correction these biases requires a combination of technical rigor and mathematical adjustment:
- Core recovery monitoring: to correct for recovery-related bias, geologists record the Total Core Recovery (TCR). If recovery falls below a certain threshold (e.g., 90-95%), the data point may be excluded from resource estimation or weighted differently. Using triple-tube drilling barrels can also physically minimize core loss in friable zones.
- The Terzaghi correction: to address orientation bias, the Terzaghi correction is applied to structural data. This involves weighting each intersection by the inverse of the cosine of the angle of intersection, effectively “boosting” the representation of features that were less likely to be hit (Fowler, 2013).
- Standardized sampling protocols: to eliminate extraction errors, companies implement strict Quality Assurance and Quality Control (QA/QC) protocols. This includes using automated core saws to ensure perfectly centered cuts and the insertion of “blinds” and “duplicates” to detect human-induced bias during the splitting and logging process.
References
Dominy, S., Glass, H., O’Connor, L., Lam, C., & Purevgerel, S. (2019). Integrating the Theory of Sampling into Underground Mine Grade Control Strategies: Case Studies from Gold Operations. Minerals, 9(4), 238. https://doi.org/10.3390/min9040238
Fowler, M. (2013). Structural data bias in the digital age. Proceedings of the 2013 International Symposium on Slope Stability in Open Pit Mining and Civil Engineering, 219-225. https://doi.org/10.36487/acg_rep/1308_09_fowler
