Designing a reliable block model for nickel deposits starts at the exploration drilling stage. The way block size and geometry are chosen must be consistent with sampling, geology, and the future mining method; otherwise, the resource will end up with low confidence and poor predictability.

Below is a practical, step‑by‑step guide for exploration geologists planning core drilling and building a nickel block model that supports robust resource classification.

1. Start from Sampling: What Does Your Core Actually Resolve?

Nickel laterite and many nickel sulphide projects use downhole sampling at about 1 m intervals, sometimes finer in transition zones (Zaitouny et al., 2021). This sampling interval sets a natural limit for vertical block height:

  • If core samples are 1 m, choose 1–2 m block height.
  • If samples are 2 m, choose 2–3 m block height (avoid blocks thinner than samples).

Using much thicker vertical blocks (e.g., 5–10 m) will mix very different horizons (limonite, saprolite, bedrock) and dilute grades and domain boundaries, reducing selectivity and adding uncertainty in the vertical direction (Zaitouny et al., 2021; Pahlawantika, Yatini and Wicaksono, 2025).

Hyperspectral and XRF core scanning can provide sub‑meter mineralogical/chemical data, but most drill databases still report composited samples at 1 m or similar (Zaitouny et al., 2021; De La Rosa et al., 2022; Harris et al., 2024). For early models, stick to the composited interval for block height; use the high‑resolution scan data mainly to refine geological domaining and contacts, not to justify unrealistically thin blocks.

2. Define Geological Domains First, Blocks Second

Modern practice is clear: domaining before estimation. For nickel laterite, key domains usually include:

  • Limonite zone
  • Saprolite zone
  • Saprock / transition
  • Fresh ultramafic bedrock

Objective boundary detection from hyperspectral or multivariate drill‑core data is powerful for this stage. Methods such as quadrant scan on hyperspectral data can automatically highlight downhole transitions that correspond to mineralogical and lithological boundaries in nickel laterites (Zaitouny et al., 2021). Multi‑sensor and hyperspectral workflows then convert continuous mineralogical curves into vertically coherent geological units suitable for 3D modelling (De La Rosa et al., 2021; De La Rosa et al., 2022).

Modern practice is clear, domaining before estimation. For nickel laterite, key domains usually include Limonite zone, Saprolite zone Saprock / transition, and Fresh ultramafic bedrock.

Automated domaining reduces subjective logging bias and makes the contact surfaces more consistent, which is critical for later block modelling and for avoiding artificial smearing of grade across true boundaries (Fresia et al., 2017; Zaitouny et al., 2021; De La Rosa et al., 2022).

3. Link Drill Spacing to Horizontal Block Size

Horizontal block dimensions must be linked to the planned or achieved drill spacing:

  • As a rule of thumb, set block length and width to about ½–¼ of nominal drill spacing.
  • For 50 × 50 m drill spacing (common for early nickel laterite work), a starting block size of 12.5 × 12.5 m is reasonable.
  • For tighter 25 × 25 m infill drilling, blocks of about 6.25 × 6.25 m can be used in detailed models.

Choosing blocks as large as the drill spacing means each block is often supported by one or zero holes, giving poor local precision and low confidence. Choosing blocks much smaller than ¼ spacing creates many poorly informed blocks, leading to unstable estimates and misleading apparent precision.

Studies in other commodities show similar behaviour: larger blocks smooth grades and increase dilution, while very small blocks become noisy and operationally impractical. An intermediate block size tied to the drilling grid balances model accuracy with mining feasibility (Mussin et al., 2025).

4. Orient and Shape the Block Model to Match the Ore

Even for laterites, the ore is not perfectly horizontal everywhere. For nickel sulphides, the geometry can be even more complex, controlled by structures, intrusions or plunging shoots (Cowan, 2020). To avoid misrepresentation of tonnage and grade:

  • Align model axes or sub‑cells with the overall strike and dip of main ore zones where possible.
  • In structurally controlled deposits, use 3D structural interpretation from drill data to understand first‑order conduits and geometry, then build domains and blocks that respect these orientations (Cowan, 2020).
  • Use sub‑blocking along sharp contacts (limonite–saprolite; ore–waste) so blocks can follow irregular surfaces without creating unrealistic “staircasing” or thick transition bands.

This structural/geometric discipline is essential; if blocks are arbitrarily aligned to a global grid, they can slice across true geological boundaries, blending ore and waste, which lowers the confidence of local grade estimates and complicates reconciliation (Cowan, 2020; Mussin et al., 2025).

5. Demonstration: Simple Design for a Laterite Nickel Project

Suppose a hill‑top laterite nickel project with the following early design:

  • Planned scout grid: 100 × 100 m RC/chip drilling, then infill to 50 × 50 m with core in higher‑potential areas.
  • Downhole sampling: 1 m intervals in limonite and saprolite, 2 m in saprock and bedrock.
  • Mining concept: open pit with 2–3 m mining benches, selective mining to separate limonite and saprolite products.

A practical block model setup for the first resource estimate could be:

  • Domains (3D solids): limonite, saprolite, saprock, bedrock, constructed using logged lithology plus hyperspectral‑based boundary detection where available (Zaitouny et al., 2021; De La Rosa et al., 2022; Pahlawantika, Yatini and Wicaksono, 2025).
  • Block size in X and Y: 12.5 × 12.5 m, derived as ¼ of 50 m infill spacing.
  • Block size in Z: 1 m in laterite profile (limonite + saprolite), 2 m in saprock/bedrock to match coarser sampling.
  • Sub‑blocking: to 3.125 × 3.125 × 0.5 m near domain boundaries and topography to respect contacts and surface.

With this design:

  • Most resource‑reporting “parent” blocks intersected by at least two holes in infill areas, reducing extrapolation distances.
  • Vertical resolution is high enough to separate thin low‑grade or waste layers from pay zones, important for later pit optimisation.
  • The number of blocks remains manageable for estimation and scheduling.

If later infill drilling closes spacing to 25 × 25 m in the central pit, the model can either:

  • Keep 12.5 × 12.5 m parents and benefit from better support (more samples per block), or
  • Locally refine to 6.25 × 6.25 m blocks in high‑grade or structurally complex zones, using variable block size concepts (Mussin et al., 2025).
For 50 × 50 m drill spacing (common for early nickel laterite work), a starting block size of 12.5 × 12.5 m is reasonable.

6. Using Geostatistics and Machine Learning to Check Block Design

Even at exploration stage, a few checks help confirm whether chosen block dimensions make sense:

  • Variography: Estimate horizontal and vertical ranges of nickel grade continuity within each domain. Block dimensions should be smaller than, or at most similar to, these ranges; very large blocks relative to range overly smooth the model.
  • Cross‑validation and swath plots: After estimation (e.g., Inverse Distance or Kriging), compare composites and block grades along sections and swaths to ensure there is no systematic bias and that trends are preserved (Bargawa, 2022; Mussin et al., 2025).
  • AI‑assisted domaining: Clustering methods and Gaussian mixture models on augmented 3D grids can improve spatial coherence of estimation domains, especially in irregular or weakly structured materials such as tailings and residues (Madani and Sabanov, 2025). Similar ideas can support better geometallurgical domaining in laterites.

Hyperspectral and multi‑sensor drill‑core workflows provide additional mineralogical variables for domaining and can be segmented at multiple scales to produce hierarchical geological units that are then modeled in 3D with standard block approaches (De La Rosa et al., 2021; De La Rosa et al., 2022).

7. Relating Block Design to Estimation Method and Resource Confidence

Estimation methods (e.g., Inverse Distance, Nearest Neighbour, Kriging) respond differently to block size and data spacing:

  • For nickel laterites, some studies have found Inverse Distance weighting performs well in limonite and Nearest Neighbour better in saprolite, while Ordinary Kriging is not always superior (Bargawa, 2022).
  • Coarse blocks combined with highly smoothing methods may hide local high‑grade features; very small blocks with sparse data can produce artificial high variability and unrealistic local extremes.

For resource classification (Measured, Indicated, Inferred):

  • Define clear criteria that combine drill spacing, number of informing samples per block, variogram ranges, and estimation variance.
  • Larger, well‑informed blocks near many holes can have higher confidence than tiny blocks inferred from single drillholes.
  • Explicitly record areas where domain boundaries, resistivity‑derived contacts, or geometallurgical units are uncertain, and keep these as lower‑confidence resources until additional drilling or geophysical surveys (e.g., ERT integration to map laterite–bedrock boundaries) reduce uncertainty (Pahlawantika, Yatini and Wicaksono, 2025).

8. Practical Tips for Exploration Geologists

When planning drilling with future block modelling in mind:

  1. Choose sampling intervals that match likely bench heights and future blocks (e.g., 1 m).
  2. Design drilling patterns (50 × 50 m, then 25 × 25 m where needed) with the idea that block size will be ½–¼ of that spacing.
  3. Use objective tools (hyperspectral, XRF scanning, automated domaining) to sharpen geological boundaries and reduce logging subjectivity (Zaitouny et al., 2021; De La Rosa et al., 2022; Harris et al., 2024).
  4. Build domains and then choose block sizes inside each domain that respect its internal continuity and intended mining selectivity.
  5. Test several candidate block sizes on a pilot area; compare grade–tonnage curves, dilution, and visual fit with geology. Select the size that best balances selectivity, stability and computational and mining practicality (Mussin et al., 2025).
  6. Document the rationale (drill spacing, sampling, variogram, mining method) for every block dimension and orientation decision; this makes audits easier and prevents ad‑hoc changes that undermine confidence.

A block model built this way—grounded in sampling, domaining, structural understanding, and practical mining geometry—will provide a higher‑confidence nickel resource, more predictable reconciliation, and a strong foundation for later mine planning and geometallurgical work.

References

Madani, N., & Sabanov, S., 2025. AI-enhanced clustering of mine tailings using Geostatistical data augmentation and Gaussian mixture models. Scientific Reports, 15. https://doi.org/10.1038/s41598-025-22625-8

Fresia, B., Ross, P., Gloaguen, E., & Bourke, A., 2017. Lithological discrimination based on statistical analysis of multi-sensor drill core logging data in the Matagami VMS district, Quebec, Canada.. Ore Geology Reviews, 80, pp. 552-563. https://doi.org/10.1016/j.oregeorev.2016.07.019

Zaitouny, A., Ramanaidou, E., Hill, J., Walker, D., & Small, M., 2021. Objective Domain Boundaries Detection in New Caledonian Nickel Laterite from Spectra Using Quadrant Scan. Minerals. https://doi.org/10.3390/min12010049

De La Rosa, R., Khodadadzadeh, M., Tusa, L., Kirsch, M., Gisbert, G., Tornos, F., Tolosana-Delgado, R., & Gloaguen, R., 2021. Mineral quantification at deposit scale using drill-core hyperspectral data: a case study in the Iberian Pyrite Belt. Ore Geology Reviews. https://doi.org/10.1016/j.oregeorev.2021.104514

De La Rosa, R., Tolosana-Delgado, R., Kirsch, M., & Gloaguen, R., 2022. Automated Multi-Scale and Multivariate Geological Logging from Drill-Core Hyperspectral Data. Remote. Sens., 14, pp. 2676. https://doi.org/10.3390/rs14112676

Bargawa, W., 2022. THE PERFORMANCE OF ESTIMATION TECHNIQUES FOR NICKEL LATERITE RESOURCE MODELING. Jurnal Teknologi. https://doi.org/10.11113/jurnalteknologi.v84.17560

Cowan, E., 2020. Deposit-scale structural architecture of the Sigma-Lamaque gold deposit, Canada—insights from a newly proposed 3D method for assessing structural controls from drill hole data. Mineralium Deposita, 55, pp. 217-240. https://doi.org/10.1007/s00126-019-00949-6

Pahlawantika, I., Yatini, Y., & Wicaksono, M., 2025. Integration of Electrical Resistivity Tomography and Borehole Data for Mapping Laterite-Bedrock Boundaries in a Nickel Deposit, ‘PHO’ Block, Southeast Sulawesi. INDONESIAN JOURNAL OF APPLIED PHYSICS. https://doi.org/10.13057/ijap.v15i1.93950

Mussin, R., Yachsishin, M., Golik, A., & Akhmatnurov, D., 2025. Block modeling reserves estimation. Kompleksnoe Ispolzovanie Mineralnogo Syra = Complex Use of Mineral Resources. https://doi.org/10.31643/2026/6445.44

Harris, A., Finn, D., MacCorquodale, F., Ravella, M., Clarke, W., Krneta, S., Battig, E., & Maguire, S., 2024. Empowering Geologists in the Exploration Process— Maximizing Data Use from Enabling Scanning Technologies. SEG Discovery. https://doi.org/10.5382/segnews.2024-136.fea-01