C3 · Publication Volume 13

Exploration Geochemistry

Follows the complete geochemical signal chain from source and dispersion to sampling, analysis and targeting.

Purpose and boundary of this book

Exploration geochemistry turns material observations into geological tests. The observable value in a sample is not a direct reading of an orebody, source rock or mineralising process. It is the end of a signal chain that includes element release, physical or chemical dispersion, selection of a sampling medium, sample support, preparation, extraction, measurement, quality control, data treatment and geological interpretation. Every link can strengthen, weaken, displace, censor or imitate the signal of interest.

This book teaches how to design, audit and interpret that chain. It covers rock, soil, regolith, stream sediment, water and biological media; orientation and production sampling; laboratory preparation and measurement; detection limits; quality-control materials; background populations; element associations; spatial effects; regolith transport; process-based targeting; and reproducible data pipelines. The goal is not to maximise the number of visually impressive anomalies. It is to identify which observations are fit to test a stated geological hypothesis, which alternatives remain viable and what action the evidence can justify.

The tutorial does not prescribe a universal sampling density, analytical package, anomaly threshold or target score. Those choices depend on the decision, expected footprint, medium, landscape process, element behaviour, concentration range and required uncertainty. It does not provide legal, environmental, safety, investment, resource-estimation or operational approval. A geochemical anomaly is an observation requiring explanation, not proof of mineralisation or economic value.

General and institution-neutral scope

This is a general, institution-neutral tutorial. It has no affiliation with, sponsorship by, endorsement from or curriculum relationship to any company or individual. It is not written for a named operator, laboratory, consultancy, university, government programme, software product, licence, mine, property or private dataset. Every unnamed locality, sample, batch, concentration and target in an example is synthetic teaching material.

People, public agencies, standards organisations, journals and repositories named in source notes identify technical sources only. A citation does not make any person or organisation the author, publisher, sponsor, partner, endorser, scientific authority or curriculum subject of this tutorial. The website that hosts the material is only a delivery surface. It is not presented as the tutorial's author, publisher, sponsor, provider, scientific authority or curriculum subject.

This boundary matters scientifically. Institutional reputation cannot repair poor sample support, an undocumented digestion, a contaminated blank, an unresolved unit conversion or an invalid spatial comparison. Conversely, a small or unfamiliar source can provide useful evidence when identity, method, uncertainty and lineage are complete. Conclusions are judged by their observable support and reproducible reasoning.

The complete geochemical signal chain

Use the following causal sequence throughout the book:

  1. A geological process creates a source distribution and mineralogical residence for elements.
  2. Weathering, erosion, dissolution, precipitation, biological uptake and mechanical transport release or redistribute material.
  3. The chosen medium integrates a spatial area, depth interval, grain-size fraction, time interval or biological pathway.
  4. Field design selects locations and sample supports from a target population under access and safety constraints.
  5. Collection, drying, sieving, crushing, grinding, splitting and storage alter the physical sample while potentially introducing loss or contamination.
  6. Extraction or digestion defines which fraction of each element becomes available for measurement.
  7. Instrument response, calibration, interference, detection limits and dilution define the reported result.
  8. Quality-control observations test contamination, accuracy, precision, mix-up and drift at stated stages.
  9. Data processing manages identifiers, units, censored values, transformations, spatial support and provenance.
  10. Geological interpretation compares process-based hypotheses and proposes a discriminating next test.

A weak link limits the conclusion. Excellent instrument precision cannot make an unrepresentative field sample representative. Dense sampling cannot make a poorly selected medium sensitive to the intended process. A smooth map cannot turn censored or mixed-support values into exact measurements. The audit therefore follows the chain in both directions: forward from hypothesis to predicted observation, and backward from a plotted value to the material and transformations that created it.

Learning outcomes

After completing the book, a learner should be able to:

  • define a decision, target population, measurand, medium and sample support before collection;
  • distinguish primary, secondary, mechanical, chemical, hydromorphic and biological dispersion processes;
  • select media by sensitivity, representativity, availability, stability, contamination risk and interpretive scale;
  • design an orientation survey and use it to set spacing, support, preparation, analysis and quality-control requirements;
  • explain how total, near-total, selective and operationally defined extractions change the meaning of a result;
  • preserve detection qualifiers and analyse censored values without treating substituted numbers as observations;
  • diagnose contamination, bias, imprecision, drift, carry-over, mislabelling and batch effects with staged control samples;
  • define background and anomaly within defensible geological and surface-process domains;
  • use element associations and log-ratios without ignoring closure, zeros, detection limits or shared denominators;
  • compare spatial values only after checking coordinate reference, support, density, domain and interpolation assumptions;
  • separate residual, transported and externally introduced signals in regolith and drainage settings;
  • build a process-based target interpretation with alternatives, negative evidence and explicit uncertainty; and
  • reconcile sample identities, batches, methods, units, qualifiers, versions and decisions in an auditable pipeline.

Prerequisites and notation

The concept map places this book after the quantitative foundations in A1, landscape and regolith reasoning in B4, geochemical process foundations in B5 and the staged exploration logic in C2. C1 provides useful mineral-system context. Learners should already be able to work with units, ratios, logarithms, simple probability, maps, coordinate systems, mass balance and competing hypotheses.

Concentration is written as mass of analyte per mass or volume of sample. Units are never inferred from the number alone. Let x_{i,e,m} denote the reported value for sample i, element e and method m; q_i is a qualifier such as below a reporting limit; s_i describes sample support; and d_i identifies the geological or process domain. A result record is incomplete unless the value, unit, qualifier, method, preparation, support and identity can be linked.

For positive parts x_1,\ldots,x_D of a composition, a log-ratio such as \ln(x_a/x_b) compares relative abundance. It is undefined when either part is a true zero and uncertain when a part is censored. Distances and interpolations depend on coordinates and support. Precision describes agreement under stated conditions; accuracy concerns closeness to an accepted reference. Neither term is a synonym for geological representativity.

How to use the diagrams and synthetic cases

Each lesson contains one purpose-built diagram. Read arrows as causal or audit relationships, not as proof that every project follows a single linear workflow. A feedback arrow means that an observation can trigger redesign; it does not permit silent retrofitting of the original hypothesis. Boxes marked as gates require a documented pass, fail or indeterminate outcome.

All unreferenced datasets are synthetic. Their values are chosen to expose calculations and failure modes, not to represent expected concentrations, background ranges, recoveries, detection limits, sampling density or exploration success. Units are stated where necessary, and arbitrary signal units are used where a real-world concentration would imply unjustified transferability. A complete worked answer contains the calculation, process interpretation, decision implication, alternative explanation and remaining uncertainty.

Reproducible geochemical records

Treat every released table and map as a view of immutable observations plus declared transformations. Preserve original laboratory fields, qualifiers and method codes. Do not overwrite a value after unit conversion, censoring treatment, coordinate correction, batch exclusion or reanalysis. Instead, create a derived field with a transformation identifier, software-independent rule, input version, operator role, timestamp and reason.

At minimum, the record should connect project-neutral sample identifiers; parent–child relationships; field coordinates and coordinate reference; medium, horizon, depth, fraction and mass; collection and custody events; preparation and analytical batches; methods and limits; results and qualifiers; quality-control roles; validation decisions; transformation history; dataset versions; and interpretive outputs. Access controls may protect sensitive coordinates or operational details, but they must not erase the lineage required by an authorised reviewer.

Assessment and completion standard

Completion requires more than reading. The learner should submit a compact synthetic survey package containing:

  1. a decision statement and two or more geological or observation-process hypotheses;
  2. a medium and support decision justified by a signal-chain model;
  3. an orientation design with acceptance and redesign rules;
  4. field and laboratory quality-control placement tied to failure modes;
  5. a data dictionary that preserves units, methods, limits and qualifiers;
  6. a censoring and compositional-data treatment justified by sensitivity analysis;
  7. domain-specific background and anomaly definitions;
  8. a support-aware spatial interpretation with unobserved areas visible;
  9. a target recommendation that includes negative evidence and an alternative explanation; and
  10. an audit report reconciling samples, batches, transformations, exclusions and released outputs.

The work passes when another reviewer can reconstruct why each observation exists, what material and process it represents, which transformations were applied, which quality tests passed, how uncertainty affects the interpretation and which result would change the proposed action. A colourful anomaly map without that chain does not meet the standard.

Core sources