🎓 Lesson 2
D2
Understanding Geological vs. Metallurgical Variability
Geological variability is how rock properties like hardness and structure change naturally across a deposit, while metallurgical variability is how the valuable mineral content and recovery behavior change from one part of the ore to another.
🎯 Learning Objectives
- ✓ Analyze drill core assay and geotechnical data to distinguish geological from metallurgical variability sources
- ✓ Explain how geological variability (e.g., joint spacing, RQD) affects fragmentation quality and its downstream impact on comminution energy
- ✓ Apply variogram analysis to quantify spatial continuity of grade vs. rock strength and interpret implications for selective mining and blending strategy
- ✓ Design a sampling and characterization protocol that captures both geological and metallurgical domains for process integration
📖 Why This Matters
A high-grade ore zone may fragment poorly due to tight jointing—causing crusher overloads—while a lower-grade, massive dolomite unit may blast efficiently but require excessive grinding energy due to hard mineral phases. Ignoring the distinction between *why* the rock breaks (geological) and *how well the metal recovers* (metallurgical) leads to costly mismatches: oversized run-of-mine material, unstable mill feed, or unanticipated reagent consumption. This lesson bridges geology and metallurgy—the two pillars of integrated mine-to-mill optimization.
📘 Core Principles
Geological variability originates from depositional history, tectonic deformation, and alteration—manifested as changes in rock type, fracture density (RQD, Jn), uniaxial compressive strength (UCS), and elastic modulus. These govern blast energy coupling, fracture propagation, and muckpile uniformity. Metallurgical variability arises from paragenesis, supergene enrichment, and micro-scale textural controls—reflected in modal mineralogy (e.g., pyrite vs. chalcopyrite), grain liberation size, gangue mineral reactivity (e.g., clay swelling), and acid consumption. Critically, these two variabilities are often *misaligned*: a geologically homogeneous zone can host extreme grade variability (e.g., vein-hosted gold), and vice versa. Process integration requires mapping and modeling them *separately*, then co-simulating their combined effect on throughput, recovery, and cost.
📐 Variogram Ratio (VR) for Domain Separation
The Variogram Ratio quantifies whether grade and rock strength exhibit similar spatial correlation structures—a prerequisite for using geological proxies (e.g., P-wave velocity) to predict metallurgical response. A VR ≈ 1 indicates strong alignment; VR ≫ 1 implies grade is more localized than rock strength, requiring independent sampling.
Variogram Ratio (VR)
VR = Range_{metallurgical} / Range_{geological}Quantifies mismatch in spatial continuity between metallurgical and geological properties; used to assess validity of geological proxies for metallurgical prediction.
Variables:
| Symbol | Name | Unit | Description |
|---|---|---|---|
| Range_{metallurgical} | Spatial correlation range of metallurgical property | m | Distance beyond which grade or recovery shows no spatial autocorrelation (from semivariogram modeling) |
| Range_{geological} | Spatial correlation range of geological property | m | Distance beyond which rock strength or lithology shows no spatial autocorrelation |
Typical Ranges:
Well-aligned porphyry copper system: 0.8 - 1.2
Vein-hosted gold with structural control: 3.0 - 8.0
💡 Worked Example
Problem: From 200 diamond drill holes at the Copper Ridge deposit: experimental semivariogram range for Cu grade = 42 m; range for UCS = 185 m. Nugget/sill ratio for grade = 0.32; for UCS = 0.18.
1.
Step 1: Extract practical ranges (distance where semivariogram reaches 95% of sill) — Cu grade: 42 m, UCS: 185 m.
2.
Step 2: Compute VR = Range(UCS) / Range(Cu grade) = 185 / 42 = 4.4.
3.
Step 3: Interpret: VR > 3 indicates grade is highly localized relative to rock strength—geological domain boundaries cannot reliably proxy metallurgical domains.
Answer:
The result is VR = 4.4, which exceeds the threshold of 3.0, confirming that grade and rock strength have dissimilar spatial structures and must be modeled independently.
🏗️ Real-World Application
At Newmont’s Boddington Gold Mine (Western Australia), initial blast designs used lithological units (granite vs. altered schist) as proxies for both fragmentation and leach kinetics. However, metallurgical testwork revealed that within the same granite unit, gold was hosted in coarse electrum grains in some zones (requiring finer crushing) and fine native gold in others (prone to preg-robbing). Geological mapping alone failed to predict this. The solution: integrate QEMSCAN® mineral liberation data with geotechnical logging to define *dual-domain blocks*—geological domains for blast design and metallurgical domains for ROM pad management—reducing cyanide consumption by 18% and improving 72-hr recovery by 6.2%.