Core, Channel and Production Sampling

Core, channel and production samples provide evidence at different positions in the mining process. Core samples preserve intervals from drill holes; channels sample defined paths across exposed material; production sampling characterizes material being drilled, moved, stored or processed. Their results are comparable only after considering what each sample represents and how collection and handling may have changed it.^[1][2]

The objective is not to identify one universally superior sampling method. It is to select and verify a method suited to the material, access conditions and decision, while retaining enough information to interpret the result at its actual spatial or production scale.

Establishing the Sampling Support

Before collection, define the material boundary, property and required reporting basis. A narrow core interval, a face channel and a composite from a moving stream do not have the same support. Their dimensions, orientation, distribution and time of collection influence how variability is averaged. CIM's estimation guidance explicitly considers these differences when data are combined for estimation.^[3]

For example, a channel may characterize exposed ground before mining, while a plant-feed sample represents broken material after extraction, handling and possible blending. A difference between their grades may involve dilution, losses, stockpile movements or different represented volumes. It cannot be assigned to analytical error without first reconciling those boundaries.

The collection record should therefore distinguish intended support from recovered material. Missing intervals, failed collections and inaccessible areas are limitations to document, not values to fill with assumed zero grades.

Sampling Drill Core

Core sampling links assay intervals to geological logging, recovery records and a surveyed hole. Sample boundaries can follow geological contacts or a justified interval convention. CIM's exploration guidance recommends considering continuous coverage through anticipated mineralization and adjoining material at discovery and delineation stages, while recognizing geological boundaries. This is a context-specific approach, not a universal sample length.^[1]

The sampled fraction also needs definition. Half-core, quarter-core and whole-core submissions provide different amounts of material and leave different archives. The choice balances analytical representativeness against future geological examination and other tests. Recording the fraction and the cutting or splitting convention is essential; a change halfway through a campaign should not disappear into a common method code.

A visually symmetrical division does not guarantee equal mineral content. A vein, coarse mineral grain or heterogeneous band may fall mainly into one portion. Consequently, paired core fractions can differ even with competent laboratory preparation. Such pairs investigate variability under the particular division and geology; they are not identical materials whose assays must coincide.^[4]

Recovery adds another qualification. The assay describes the submitted material, whereas the intended geological interval may include missing material. If loss is selective, applying the recovered grade to the complete interval can introduce bias. Retained core, photographs, interval records and recovery observations help evaluate that risk, but no laboratory repeat can reconstruct the unknown composition of material never recovered.^[1]

Channels and Exposed-Face Samples

A channel is intended to sample a defined path through exposed geology. Its position, direction, sampled width and extraction depth establish what material is included. Collection that preferentially removes soft or friable components can alter composition even when the marked line is correct. Minnitt's South African review describes this problem for particular reef settings; those examples demonstrate a mechanism rather than a universal correction factor.^[2]

The observed exposure may also be weathered, coated or disturbed. The relevant question is whether those conditions belong to the target being characterized or obscure the underlying material of interest. Any removal or exclusion changes the sampling boundary and should have a documented geological reason rather than merely improving the sample's appearance.

Channel length is measured along the collection path. It is not automatically the true thickness of the intersected body, and a chemically valid composite cannot resolve that geometrical distinction. The interpretation needs exposure geometry and structural observations in addition to assays.

Safe access is a design constraint. Where the planned exposure cannot be sampled appropriately, the record should identify the coverage gap and its implications. A convenient selective specimen can still be useful for mineral identification, but it should not be relabelled as a representative channel sample.

Sampling Broken Material and Production Streams

Production sampling may involve drill cuttings, blast-hole material, stockpiles, conveyor streams or process slurries. Each requires its own assessment of segregation, recovery, contamination and timing. Drill cuttings, for example, can be displaced or mixed during collection, while a stockpile surface may have a different particle-size distribution from its interior. A familiar method name does not establish unbiased recovery.^[2]

For a moving stream, the collection system should be assessed against the stream actually presented to it. Missing part of the cross-section, overflowing a cutter or losing fine material can exclude particular constituents. Inspection and performance testing are therefore needed in addition to a nominal equipment specification. An automated sampler can repeat a biased collection just as consistently as a manual procedure.^[2]

Time-based sampling is not inherently incompatible with a mass-weighted lot grade. A 2014 IEA Clean Coal Centre report describes time-basis sampling with increment mass proportional to the instantaneous flow, contrasted with mass-basis sampling using uniform increment masses. This coal example illustrates the need to specify both timing and mass: subsequently taking equal masses from unequal time-based increments can discard the intended weighting. It is not a universal sampler specification.^[5]

For slurries, also identify whether the property concerns dry solids, liquid or the combined stream. A production average needs a weighting basis appropriate to that defined material.

A production record should identify sampling periods, operating interruptions, routing changes and relevant inventory boundaries. These details allow an analyst to determine whether a composite covers the intended production or only the periods when sampling equipment happened to be available.

Composites and Comparisons

Physical composites combine material before analysis; calculated composites combine separate results. Neither approach removes the need to preserve original identities and weights. A composite across contrasting domains may serve a blending question while being unsuitable for defining the individual geological domains. CIM recommends relating compositing to support, geological boundaries and the intended use.^[3]

Where mass grades are combined, weights should reflect represented mass. Equating length-weighted geological composites with mass weighting requires appropriate cross-sectional and density assumptions. Neither convention removes bias from selective material loss. The masses in sample bags are not automatically the tonnages those bags represent.^[3][4]

Cross-method comparisons need a matched target. Comparing a new channel with an old core interval can be informative, but differences in position and support remain. Comparing a production lot with a resource model additionally requires the mined boundary and material movements. Laboratory agreement alone cannot settle those comparisons.

Quality Checks and Traceable Handover

Quality controls should be inserted where they can test the suspected errors. Field repeats investigate the collection stages repeated; preparation duplicates investigate the splitting, size reduction and analysis actually repeated; repeats from an assay pulp (pulverised laboratory powder) examine a narrower part of the chain. Shared upstream errors remain shared, even if the duplicate results agree closely.^[4]

Independent laboratory checks can investigate analytical differences, but they need matched decomposition, concentration ranges and reporting bases. CIM's Canadian guidance places these checks alongside blanks, reference materials and duplicates rather than treating any one as complete validation.^[3]

The final handover includes location or lot identity, interval or period, collection method, recovered fraction, preparation instructions and custody records. Unexpected results should lead back through these records to the actual material and procedure. That traceability allows a decision to distinguish geological variation from sampling limitations, preparation problems or analytical differences instead of treating every disagreement as a laboratory failure.^[1]