The Link Between Sampling and Resource Estimation

 The Link Between Sampling and Resource Estimation

---

Why Sampling Matters for Resource Estimation

1. Data Foundation

   • Resource estimation relies on the data collected from the deposit. Sampling (via drilling, trenches, channel samples, etc.) is how you collect “hard data” about grade (metal content), density, lithology, and other characteristics. 

   • If samples are not representative, the model built from them may misrepresent the true distribution of the mineral.

2. Spatial Continuity & Variability

   • Geological bodies are rarely homogeneous. There’s spatial variability (nugget, structure, geologic domains), and sampling determines how well that variability can be characterized. 

   • Geostatistical tools (e.g., variogram) use sampled data to measure how grade or other properties vary with distance. These parameters feed directly into estimation techniques like kriging.

3. Uncertainty Quantification

   • Sampling introduces several sources of uncertainty:

     • Sample collection: errors in how samples are taken (bias, contamination, volume) 

     • Analysis: laboratory measurement error, detection limits, missing data, outliers 

     • Spatial positioning: where drill holes are located, spacing, dip, and orientation impact how well the sample network captures the deposit’s geometry. 

   • These uncertainties propagate into the estimated model. More or better-quality samples generally reduce estimation uncertainty. 

   • The classification of resources (measured, indicated, inferred) often reflects the level of uncertainty, which is tied to sampling confidence. 

4. Modeling & Estimation Methods

   • Once samples are collected and analyzed, estimation methods (e.g., kriging, conditional simulation) are used to predict values in unsampled blocks. The reliability of these predictions depends on how well the sample data represent the deposit. 

   • For example, kriging variance (a measure of estimation error) depends on sample location, density, and sample support (size/shape of sample) — so sampling design directly affects model uncertainty. 

   • More advanced approaches (like geostatistical simulation) can quantify the range of possible outcomes (i.e., uncertainty distribution), but still depend on sample data. 

5. Regulatory & Reporting Standards

   • Industry guidelines emphasize proper sampling protocols. For example, the CIM “Best Practice Guidelines” list sampling theory, sample preparation, QA/QC, etc., as critical components of resource estimation.

   • Resource reporting standards (e.g., JORC, NI 43-101) require that sampling is done in a way that supports the claimed classification and confidence level.

6. Economic Decisions

   • Resource estimates based on poor sampling can mislead economic evaluations, mine planning, and investment. Overestimating grade or tonnage can lead to bad decisions; underestimating can lead to missed opportunities. 

   • Quantifying uncertainty (which comes from sampling) helps in risk assessment, capital allocation, and target prioritization. 

---

Key Risks & Challenges

• Non-representative samples: If sampling is biased (e.g., only high-grade zones are drilled), resource models will overestimate.

• Insufficient sample density: Sparse drilling may not capture variability; interpolation will be less reliable.

• Poor QA/QC: Without checks on lab data, handling contamination, or sample preparation errors, the input data may be flawed.

• High nugget effect: If the deposit has high short-scale variability (nugget), even closely spaced samples might not capture real variability, making estimation difficult. 

• Incorrect domain modeling: If geological domains (e.g., different lithologies or mineralization zones) are not correctly defined based on sample data, the estimation will misallocate grades.

---

How the Link Is Managed in Practice

1. Sampling Strategy Design

   • Geologists design sampling campaigns (drill spacing, sample size, orientation) based on geological understanding to ensure representativeness.

   • Use of systematic sampling, stratified sampling, or other methods depending on deposit style. 

2. Quality Assurance / Quality Control (QA/QC)

   • Inserting blanks, duplicates, standards into sample streams to check for lab and handling errors.

   • Regular audits of sampling processes to avoid bias or contamination.

3. Geostatistical Analysis

   • Variogram modeling to understand spatial continuity. 

   • Using geostatistical estimation (kriging) that weights samples based on spatial correlation to produce unbiased estimates. 

   • Running simulation (conditional simulation) to assess uncertainty and generate multiple realizations of the resource. 

4. Validation & Classification

   • Validate estimated models against “hard data” (e.g., measure model performance, cross-validation).

   • Use classification (e.g., inferred / indicated / measured) to reflect confidence based on data density, variability, and estimation error.

   • Report uncertainty explicitly (as required in some standards) to inform stakeholders. 

---

More information visit:

https://ko-fi.com/orezone

https://t.me/OreZone