Block Model Estimation: From Drill Holes to Mine Planning

simulator advanced ~14 min
Loading simulation...
45% ore blocks — avg ore grade 4.8 g/t at cutoff 2.5 g/t

With a mean grade of 3.5 g/t and cutoff of 2.5 g/t, approximately 45% of blocks classify as ore with an average ore grade of 4.8 g/t. The block model shows the spatial distribution of ore and waste zones.

Formula

Z*_B = Σ λᵢ Z(xᵢ) (block kriging estimate)
Ore tonnage = Σ T_B for blocks where Z*_B ≥ g_c
Metal = Σ (Z*_B × T_B) for ore blocks

From Drill Holes to Dollars

The block model is the fundamental bridge between geological data and mining economics. Every open pit and underground mine in the world relies on a block model — a three-dimensional grid where each block carries an estimated grade, tonnage, and classification. Drill hole samples, typically spaced tens to hundreds of meters apart, are the sparse input. Geostatistical estimation (kriging) fills the gaps, producing the best possible grade estimate for every block in the deposit along with its uncertainty.

Cutoff Grade Optimization

The economic cutoff grade determines which blocks are ore (worth mining and processing) and which are waste (removed but not processed). In this simulation, you can adjust the cutoff and immediately see how it reshapes the ore body — raising the cutoff shrinks the ore envelope and increases the average ore grade, while lowering it expands the envelope but dilutes the grade. The optimal cutoff depends on metal price, mining cost, processing cost, and recovery — Lane's algorithm formalizes this optimization.

Understanding the Visualization

The simulation displays a 2D cross-section of a block model. Each block is colored by its estimated grade — hot colors (red/yellow) indicate high grades, cool colors (blue) indicate low grades. The cutoff grade is shown as a contour line separating ore (opaque) from waste (translucent). Drill hole locations are marked as white dots. Notice how blocks near drill holes have sharp color contrasts (confident estimates) while blocks far from data trend toward the mean grade (regression to mean — a kriging property).

Resource Classification and Risk

Block models support the formal resource classification system used worldwide. Blocks estimated from dense, closely-spaced drilling are classified as Measured resources — the highest confidence category. Blocks estimated from moderate drill spacing are Indicated. Blocks estimated from sparse data or extrapolated are Inferred — too uncertain for economic evaluation. The transition from Inferred to Measured represents a journey from geological speculation to bankable reserves, with each step requiring more drilling, better variogram models, and lower kriging variance.

FAQ

What is a block model?

A block model divides the subsurface into a regular 3D grid of blocks (typically 5-25 m). Each block is assigned an estimated grade based on nearby drill hole samples using kriging or other geostatistical methods. The block model is the foundation of mine planning, resource estimation, and economic evaluation.

How does cutoff grade affect ore reserves?

The cutoff grade is the minimum grade at which mining is economic. Lowering the cutoff increases ore tonnage but decreases average ore grade. Raising the cutoff decreases tonnage but increases grade. The optimal cutoff maximizes net present value considering mining costs, processing costs, metal price, and recovery.

What is the support effect in block estimation?

Block grades have lower variance than point samples because kriging averages over the block volume (change of support). This means the block model has a tighter grade distribution than drill hole data — fewer extreme highs and lows. This is critical because misunderstanding support effects leads to incorrect resource classification.

How are resources classified using block models?

Mining resources are classified as Measured, Indicated, or Inferred based on kriging variance (estimation confidence). Measured resources have dense drill spacing and low kriging variance. Indicated have moderate spacing. Inferred have sparse data and high uncertainty. Only Measured and Indicated can be converted to reserves.

Sources

Embed

<iframe src="https://homo-deus.com/lab/geostatistics/block-model/embed" width="100%" height="400" frameborder="0"></iframe>
View source on GitHub