Geochemical sampling is a low-cost, early-stage exploration technique that collects natural materials — soil, stream sediment, or rock — from across a target area and analyses them for trace concentrations of metals. Because weathering and water transport spread metal signatures away from their source, a geochemical survey can flag mineralisation buried under cover or hidden beneath vegetation long before anyone drills a hole.
This guide walks through the three sample media used in almost every ground geochemical program, explains how sampling density and depth are chosen, and shows how geologists separate a genuine anomaly from ordinary background variation.
Why sample dirt and rock instead of just drilling?
Drilling is the only way to prove what’s underground, but it’s also the most expensive step in exploration, often costing hundreds of dollars per metre once mobilisation, logging, and assay costs are included. Before committing to a drill program, exploration teams need a cheap, fast way to narrow a large area down to a handful of promising targets.
Geochemical sampling fills that role. Weathering breaks down mineralised rock at surface, and the released metals disperse outward — dissolved in groundwater, adsorbed onto soil particles, carried in stream sediment, or simply exposed in outcropping rock chips. A geochemical survey measures that dispersion. A cluster of samples running well above local background for pathfinder elements like arsenic, antimony, or copper tells a geologist where to look next, whether that means tighter-spaced sampling, a geophysical survey, or, eventually, a drill hole.
The three main sample media
Soil sampling
Soil sampling collects a consistent horizon of soil across a grid or along traverse lines, on the principle that metals liberated from buried mineralisation migrate upward and outward through the soil profile and adsorb onto clay minerals and iron/manganese oxides.
Samples are usually taken from the B-horizon — the subsoil zone of accumulation beneath the organic-rich A-horizon — because it is less affected by recent organic input, more stable through time, and more likely to retain a secondary dispersion signature from depth. In practice this typically means digging to roughly 25–30 cm, though depth is adjusted to local soil development.
Grid spacing depends on the stage of exploration:
- Reconnaissance surveys covering a large regional area commonly use widely spaced lines, on the order of 200–400 m between lines and 100–400 m between samples along the line.
- Detailed follow-up grids over a defined target tighten this to roughly 50–100 m spacing in both directions, sharp enough to resolve the shape of an anomaly and guide where a drill program should focus.
Soil sampling works best in areas with a well-developed, in-situ residual soil profile. It becomes unreliable in terrain with thick transported cover, strong slope wash, or where soils have been disturbed, because the metal signature may no longer sit above its actual source.
Stream sediment sampling
Stream sediment sampling collects material from active drainages, on the principle that a stream integrates the weathering products of its entire upstream catchment — making it an efficient way to screen a large area with relatively few samples.
Sampling normally targets the fine, active silt fraction (commonly specified as –80 mesh, i.e. material passing an 80-mesh sieve) taken from the active channel, because fine sediment concentrates the clay-adsorbed and oxide-bound metal fraction most likely to carry a hydromorphic dispersion signal. Geologists sample from natural traps where denser sediment accumulates — the inside of bends, behind boulders, in stream margins — rather than from wherever is convenient.
Because each sample represents an entire catchment, spacing is planned around drainage basin size rather than a fixed grid. Regional reconnaissance programs commonly target catchments in the range of roughly 5–50 km², sampling near the mouth of each basin, with supplemental samples added where basin sizes fall outside that range or where drainage density is uneven. Where a known deposit exists, orientation sampling both upstream and downstream — often at roughly 250–350 m spacing — helps establish how far the metal signature actually disperses down-drainage before planning the wider survey.
Stream sediment sampling is one of the most cost-effective first-pass tools in exploration, but it only works where there is an active drainage network, and a single anomalous sample only narrows the target down to “somewhere in this catchment” — follow-up soil sampling or mapping is needed to pinpoint the source.
Rock chip sampling
Rock chip sampling collects fragments of outcropping or sub-cropping rock directly, wherever bedrock is exposed at surface. Unlike soil or sediment sampling, it samples the primary source material itself rather than a secondary dispersion halo, which makes it useful for confirming mineralisation style and grade at a specific location — but it comes with a well-known bias risk.
Two collection styles are used, and they answer different questions:
- Grab sampling takes one or a few pieces from a location without a fixed pattern. It is fast and cheap, and useful for quickly checking whether mineralisation is present at all — but because geologists naturally gravitate toward the best-looking rock, a grab sample is not a reliable estimate of average grade across a rock unit and should never be reported or treated as representative tonnage-grade data.
- Channel sampling cuts a continuous groove across a rock face, vein, or outcrop at a fixed width and spacing, collecting all material along that line rather than cherry-picking pieces. It is more systematic and far less prone to selection bias, which is why it is the preferred method for estimating true grade and continuity across a mineralised structure — though it is more labour-intensive and, done carelessly with just a hammer and chisel, can drift back toward an irregular, grab-like result if the sampler doesn’t maintain a consistent groove.
Rock chip programs are typically used to follow up a soil or stream sediment anomaly once mapping has identified the outcropping source, or to characterise a vein or structure ahead of a drill program.
Comparing the three methods
| Method | What it samples | Typical spacing/scale | Best terrain | Main limitation |
|---|---|---|---|---|
| Soil sampling | B-horizon soil, in-situ residual profile | 50–100 m (detailed); 200–400 m (reconnaissance) | Well-developed residual soil, moderate cover | Unreliable over thick transported cover or disturbed ground |
| Stream sediment sampling | Fine (–80 mesh) active channel sediment | One sample per ~5–50 km² catchment | Areas with active drainage networks | Only localises to a catchment, not a source; needs follow-up |
| Rock chip — grab | Individual outcrop/float fragments | Opportunistic, wherever exposed | Outcropping bedrock | Strong selection bias; not representative of grade |
| Rock chip — channel | Continuous groove across a face or vein | Fixed interval along the exposure | Outcropping bedrock, vein/structure exposed | Labour-intensive; requires careful, consistent cutting |
How do geologists tell a real anomaly from background noise?
Every element has a natural “background” concentration in ordinary, unmineralised rock and soil — and background is better understood as a range than a single number, since it varies with rock type, climate, and local geology. A threshold is the value that marks the upper (or lower) edge of that normal range, and any sample result that sits clearly above threshold is treated as an anomaly worth following up.
In practice, this is rarely a single clean cutoff. Geologists typically plot the full population of results — often on a cumulative probability plot — to separate the background population statistically from one or more higher populations that may represent real mineralisation, before drawing threshold lines specific to that survey and that geological setting. A single elevated sample can also just be noise (contamination, an unusual local mineral, a lab error); geologists look for coherent clusters of anomalous samples that line up with geological structure or known controls before committing follow-up budget.
Pathfinder elements — elements that occur alongside the target metal but disperse more widely or are easier to detect at low concentration, such as arsenic or antimony around some gold systems — are often tracked alongside the commodity of interest because they can extend the effective “halo” a survey can detect.
Common pitfalls in geochemical sampling
- Contamination. Sampling too close to roads, old workings, fence lines, or camp infrastructure can introduce metal contamination (galvanised wire, vehicle exhaust, old tailings) that has nothing to do with bedrock geology.
- Inconsistent sample media. Mixing soil horizons, sediment size fractions, or sample depths across a program destroys comparability between samples and can manufacture a false anomaly out of nothing more than a change in method.
- Treating grab samples as representative. As above — a handful of eye-catching grab samples is not evidence of grade; it’s evidence that mineralisation exists somewhere nearby.
- Under-recording sample metadata. Sample location, medium, horizon/depth, and collection method all need to be captured consistently, or a promising anomaly becomes impossible to properly interpret or follow up months later.
Programs that log sample metadata consistently from day one — location, medium, depth, QC type — make it far easier to trust an anomaly enough to act on it. This is exactly the kind of structured field data capture that Blue Butterfly’s browser-based logging and cloud geological database are built to support, whether the data is coming from a drill core or a geochemical sampling program.
FAQ
What’s the difference between geochemical sampling and geophysical surveying? Geochemical sampling measures the actual chemical composition of soil, sediment, or rock, directly detecting metal dispersion at or near surface. Geophysical surveying instead measures physical rock properties — magnetism, conductivity, density — that can indicate structures or alteration associated with mineralisation, without directly measuring metal content. The two are complementary and are often run together or in sequence.
Which sample medium should I choose first? It depends on terrain and scale. Stream sediment sampling is usually the fastest, cheapest way to screen a large, poorly known area with active drainages. Soil sampling follows to define anomalies in more detail once a target area is narrowed down. Rock chip sampling comes in once outcrop is identified, to confirm mineralisation style directly.
How many samples make an anomaly credible? There’s no fixed number, but geologists look for coherence rather than a single outlier: multiple adjacent or nearby samples elevated above threshold, ideally lining up with a mapped structure, lithological contact, or alteration zone, rather than one isolated high value that could be contamination or lab error.
Can geochemical sampling detect deposits under deep cover? Conventional soil and stream sediment sampling work best where cover is thin to moderate and weathering products can migrate to surface. Under thick transported cover, standard near-surface geochemistry becomes unreliable, and programs may turn to deeper sampling techniques, groundwater or gas geochemistry, or geophysics instead.
What does a “pathfinder element” actually tell you? A pathfinder is an element that travels with or ahead of the target commodity’s dispersion halo and is often easier or cheaper to detect at trace levels. A strong pathfinder anomaly doesn’t guarantee the target metal is present in economic quantity — it flags a location worth checking with more direct methods.
Sources
- U.S. Geological Survey — Geochemical Data Portal for Rock, Sediment, Soil, and Mineral Samples, United States and Territories: https://www.usgs.gov/tools/geochemical-data-portal-rock-sediment-soil-and-mineral-samples-united-states-and-territories
- U.S. Geological Survey — Procedures Manual for Stream Sediment Reconnaissance Sampling: https://pubs.usgs.gov/of/1997/ofr-97-0492/pubs/gjbx_84(78).pdf
- Geological Survey Ireland — Stream Sediments (Mineral Exploration Activities): https://www.gsi.ie/en-ie/programmes-and-projects/minerals/activities/mineral-exploration/Pages/Stream-Sediments.aspx
- ScienceDirect / GEEA — Stream sediment geochemistry in mineral exploration: a review of fine-fraction, clay-fraction, bulk leach gold, heavy mineral concentrate and indicator mineral chemistry: https://pubs.geoscienceworld.org/gsl/geea/article/23/4/geochem2022-039/624132/Stream-sediment-geochemistry-in-mineral
- ScienceDirect — Geochemical background — concept and reality (Reimann & Garrett): https://www.sciencedirect.com/science/article/abs/pii/S0048969705001889
- ScienceDirect — Determination of background and threshold in exploration geochemistry: https://www.sciencedirect.com/science/article/abs/pii/0375674274900028
- Roger Marjoribanks — Collecting Rock Chip Samples: https://rogermarjoribanks.info/collecting-rock-chip-samples/
- 911Metallurgist — Mining Geology Sampling Methods: Channel, Chips, Core: https://www.911metallurgist.com/blog/mining-geology-sampling-methods-channel-chip-core/
- SciELO — The ‘simulated chip-sample model’ as a method for quantifying error and bias in sampling thin carboniferous reef types: https://scielo.org.za/scielo.php?script=sci_arttext&pid=S2225-62532014000100008
- Rangefront — Soil Sampling Techniques in Mineral Exploration: https://rangefront.com/blog/soil-sampling-techniques/