E1 ยท Publication Volume 23

CRS Definition and Metadata

EPSG identifiers, WKT, axis order, units and area of use

Learning objectives

read a CRS as a structured object rather than a label; use authoritative identifiers without treating them as complete evidence; inspect WKT components, axes, units, extent and dimensionality; and design metadata that survives exchange.

The objective is transferable reasoning, not operation of a named product or performance of regulated survey work. Every real decision must use current applicable requirements, authorised control and competent review.

Decision context

The decision is whether a dataset has enough information to locate every coordinate unambiguously and reproduce every coordinate operation. A short identifier is useful only when it resolves to the intended version and definition. A WKT string is useful only when it is preserved and parsed consistently. The analyst compares identifier, name, structured definition, file header, numeric range, location evidence and lineage rather than trusting one field.

Metadata repair and coordinate transformation are different actions. Adding the definition that was already used changes metadata but not coordinates. Assigning a guessed definition creates a hypothesis. Reprojecting from a verified source creates new coordinate values. The audit log must distinguish all three.

Core concept

Complete CRS definition and validation layers: simplified institution-neutral teaching model
Complete CRS definition and validation layers: simplified institution-neutral teaching model

A CRS object contains a datum or reference frame, ellipsoid where applicable, coordinate system, axes, units and scope or extent. A projected CRS also contains a base geodetic CRS and coordinate-conversion method with parameters. A compound CRS joins compatible components such as horizontal and vertical references. A bound or operation-aware description may include a preferred transformation but does not remove the need to review area and accuracy.

WKT is a structured representation, not decorative text. Names are human-readable, while identifiers connect to an authoritative registry. Axis clauses, unit clauses, datum ensemble or dynamic-frame elements, reference epoch, conversion parameters and usage extent can materially change interpretation. Older and newer WKT dialects may lose information when converted.

Reference frames and metadata

The minimum inventory captures dataset identifier, geometry type, coordinate dimension, CRS identifier, full WKT or equivalent structured definition, axis order, units, horizontal datum realisation, coordinate epoch, vertical CRS, area of use, source format, software-independent operation record, definition version and checksum. Original headers remain evidence even when flawed.

Do not infer CRS solely from coordinate magnitude. Easting 500000 may be a false easting in many projected systems or a local grid value. Longitude 145 may be degrees or an unlabelled local number. Range and location tests narrow hypotheses but require independent control to select one.

Quantitative reasoning

A metadata completeness score may be used only as a checklist: C=\sum_i w_i q_i, where each q_i records whether an essential field is present and verified. It is not a spatial-accuracy metric. Missing datum or unit should be a hard failure even if many optional fields are present.

Coordinate bounds provide conditional tests. For a candidate CRS, transform the declared area-of-use boundary and test whether dataset coordinates fall within a reasonable envelope. Then compare known controls. Passing a broad range test cannot validate the candidate; failing a well-constructed range test can reject it.

Evidence and uncertainty

Evidence includes original file bytes, sidecar metadata, database schema, export settings, data-provider statement, registry snapshot, WKT, operation logs, control-point coordinates and spatial context. Preserve conflicting definitions rather than silently choosing one. The source with the most detailed string is not necessarily correct if it was attached during a later export.

Uncertainty here includes semantic uncertainty: the probability that the definition itself is wrong. A centimetre survey with uncertain CRS can have metre-scale or larger positional ambiguity. Record definition confidence separately from numerical coordinate accuracy.

Transformation and control

A CRS ingestion gate parses every definition, validates syntax, resolves identifiers, compares equivalent components, checks axes and units, checks area of use, verifies dimensions and requires an explicit vertical decision. It assigns states such as verified, internally consistent but unverified, conflicting, inferred hypothesis or rejected. Only verified sources enter automatic transformation.

Changes are non-destructive. Preserve source_crs_raw, source_coordinates and source_file_hash; add normalised_crs and interpretation_version; then create transformed coordinates with operation and target fields. A later correction can invalidate descendants through lineage without erasing history.

Interfaces and data

Files, databases and APIs encode CRS differently. The exchange contract states which element is authoritative when identifier and embedded WKT disagree, how axis order is serialised, whether coordinates follow CRS axis order or a format-specific order, how vertical components are represented and whether epoch metadata travels at dataset or feature level.

A coordinate identifier in a URL, a numeric database code and a WKT authority node can all refer to the same registry entry yet be versioned differently. Store the resolved definition snapshot or release alongside the identifier for reproducibility.

Integration checkpoint

The checkpoint passes when the learner can reconstruct a CRS from its components, detect a disagreement among identifier, WKT and numeric evidence, and explain why assigning a definition is not reprojection. At least one ambiguous dataset must remain unresolved if evidence is insufficient.

The strongest test combines semantic evidence with independent spatial control. A clean parse only proves syntactic validity; it does not prove the definition belongs to the coordinates.

Synthetic worked example

A synthetic raster header names a projected CRS, its embedded WKT uses another datum, and a sidecar identifier points to the correct zone but wrong datum generation. Pixel coordinates overlap the expected project only under two candidates. The learner preserves all three claims, checks the original export log and tests three independent controls.

The WKT is found to have been injected by a later conversion. The source identifier is repaired from documented export evidence without changing pixels; a separate reprojection product is then created. The example is synthetic and names no software or organisation.

Practice task

Audit five synthetic spatial packages: a shapefile-style bundle, a GeoTIFF-style raster, a CSV, a database table and an API response. Build a uniform CRS inventory, classify conflicts, resolve only evidence-supported cases and design a rejection message that tells a supplier exactly which fields are missing.

Submit raw metadata snapshots, normalised definitions, confidence states, validation controls and lineage links.

Common failure modes

The following failures are treated as evidence or process defects, not cosmetic issues:

  • treating an identifier as proof it belongs to the data.
  • equating CRS assignment with reprojection.
  • ignoring axis order or units.
  • dropping vertical and epoch metadata.
  • converting WKT dialects without checking information loss.
  • inferring from coordinate range alone.
  • overwriting conflicting source metadata.

For each failure, preserve the original evidence, identify its downstream reach, define a discriminating test and record whether the case is corrected, rejected or still unresolved.

Review questions

  1. Which components make a projected CRS complete?
  2. Why can valid WKT still be wrong for a dataset?
  3. How should identifier and embedded definition conflicts be handled?
  4. What is semantic coordinate uncertainty?
  5. Which records must remain immutable during metadata repair?

Answer with definitions, evidence, a calculation or test where relevant, and the condition that would reverse the conclusion. A product screenshot or unexplained code is not an answer.

Assessment artefact

The assessment artefact is a CRS metadata audit and repair package. It contains raw evidence, parsed components, registry snapshots, conflict matrix, hypothesis tests, control residuals, non-destructive repairs, transformation descendants and a machine-readable data contract. It must demonstrate that uncertainty was preserved where evidence could not identify one definition.

Sources and further reading