Domain vocabularies, ontologies and units
Scientific terminologies, concept schemes, ontologies and unit vocabularies used to describe domain entities, methods, properties, processes and quantitative values.
This resource helps researchers and data stewards identify authoritative semantic resources and apply their terms, identifiers and units in structured research metadata.
Domain vocabularies support precise scientific description, while mappings to cross-disciplinary metadata preserve discovery and interoperability across repositories, catalogues and research infrastructures.
Resource information
Resource type: Semantic resource guidance
Resource level: Scientific domain, method, instrument, property and measurement level
Primary audience: Researchers, data stewards, domain experts, metadata specialists, repository managers and research software developers
Primary uses: Dataset annotation, metadata profile development, semantic search, data integration, validation and machine-actionable exchange
Semantic resources: Controlled vocabularies, taxonomies, thesauri, ontologies, classifications and unit systems
Representation models: SKOS, OWL, RDF, JSON-LD, persistent concept URIs and machine-readable unit codes
Last reviewed: July 2026
Purpose and scope
Domain vocabularies and ontologies provide shared concepts for describing the scientific meaning of data. They may represent materials, chemical entities, organisms, environmental phenomena, research methods, instruments, variables, properties and processes.
Unit vocabularies provide unambiguous identifiers and codes for quantities, dimensions and units of measure. They allow values to be validated, compared and converted automatically.
These semantic resources extend general metadata vocabularies. They should not duplicate common elements such as resource type, contributor role, access status or licence.
Important: use a maintained community vocabulary whenever an appropriate one exists. A local term should be introduced only when the required concept is not represented adequately.
Types of domain semantic resources
Controlled vocabulary
A maintained set of authorised terms or codes used consistently within metadata and information systems.
Taxonomy or thesaurus
Organises concepts through preferred labels, synonyms and broader, narrower or related terms.
Ontology
Formally represents classes, properties, relationships and constraints within a scientific domain.
Unit vocabulary
Defines quantities, dimensions, measurement units, symbols and machine-readable codes.
Choosing the appropriate semantic model
| Semantic resource | Typical structure | Use when |
|---|---|---|
| Code list | A finite set of permitted values and codes | A metadata field accepts only a small number of clearly defined values. |
| Controlled vocabulary | Authorised terms with definitions and identifiers | Consistent terminology is required without a complex hierarchy. |
| Taxonomy | Hierarchical organisation of broader and narrower concepts | Resources must be classified and browsed by subject or type. |
| Thesaurus | Preferred terms, synonyms and semantic relationships | Indexing, multilingual search and query expansion are required. |
| Ontology | Classes, properties, axioms and formally defined relationships | Data integration, automated inference or detailed domain modelling is required. |
| Unit vocabulary | Quantities, dimensions, units, symbols and conversion relations | Numerical values must be represented and processed unambiguously. |
What domain semantic resources describe
Scientific entities
Materials, substances, samples, organisms, locations, structures and other objects of investigation.
Methods and processes
Experiments, simulations, measurements, transformations and analytical procedures.
Properties and variables
Measured or calculated characteristics, observables, parameters and quality indicators.
Quantities and units
Numerical values, dimensions, measurement units, uncertainties and conversion rules.
Elements of a vocabulary concept
| Element | Purpose |
|---|---|
| Concept URI | Provides a stable machine-readable identifier independent of the displayed wording. |
| Preferred label | Provides the authorised human-readable term in a particular language. |
| Alternative label | Records synonyms, abbreviations, spelling variants or historical terms. |
| Definition | Explains the concept and distinguishes it from related concepts. |
| Broader concept | Identifies a more general concept in the same hierarchy. |
| Narrower concept | Identifies a more specific concept in the same hierarchy. |
| Related concept | Connects a semantically associated concept outside the direct hierarchy. |
| Mapping | Relates the concept to an exact, close, broader, narrower or related concept in another vocabulary. |
| Status | Indicates whether the concept is active, proposed, deprecated or replaced. |
| Version information | Identifies the vocabulary release in which the concept was used. |
Selecting a domain vocabulary or ontology
1. Identify the concept
Determine whether the metadata describes an entity, method, process, property, variable or quantity.
2. Search community resources
Review registries, domain repositories and semantic resources used by the relevant research community.
3. Evaluate the vocabulary
Check coverage, governance, persistent identifiers, versioning, licence and machine-readable access.
4. Record the decision
Document the selected concepts, vocabulary version, local constraints and mappings.
Selection criteria
Domain relevance: the resource represents the required scientific concepts at an appropriate level of detail.
Community adoption: it is used by recognised repositories, infrastructures, projects or research communities.
Governance: maintainers, contribution procedures and responsibilities are documented.
Persistent identifiers: concepts are identified by stable and resolvable URIs.
Machine-readable access: structured downloads, APIs, SPARQL or content negotiation are supported.
Versioning: releases, deprecated terms and replacements are documented.
Definitions: concepts have precise, non-circular and understandable definitions.
Mappings: correspondences with related semantic resources are available or can be created.
Licence: reuse conditions allow implementation in the intended metadata system.
Finding domain vocabularies and ontologies
| Service | Coverage | Recommended use |
|---|---|---|
| FAIRsharing | Metadata standards, formats, terminologies, identifier schemes, databases and policies | Find community standards and examine their relationships with repositories and policies. |
| NCBO BioPortal | Biomedical and life-sciences ontologies | Search terms, inspect ontology hierarchies and access biomedical semantic resources. |
| NERC Vocabulary Server | Marine, oceanographic and environmental concept collections | Use persistent SKOS concepts for variables, instruments, platforms and marine observations. |
| MatPortal | Materials-science and engineering ontologies | Search, browse and compare semantic resources for materials data. |
Examples of domain semantic resources
| Domain | Semantic resource | Typical concepts |
|---|---|---|
| Agriculture and food | AGROVOC | Crops, organisms, farming systems, food products, processes and environmental concepts |
| Environment | GEMET | Environmental topics, policies, processes, substances, risks and geographic concepts |
| Marine science | NERC Vocabulary Server | Observed properties, instruments, platforms, sampling devices and data-processing terms |
| Chemistry and chemical biology | ChEBI | Atoms, molecules, ions, chemical substances, structural classes and biological roles |
| Molecular biology and genetics | Gene Ontology | Molecular functions, biological processes and cellular components |
| Materials science | EMMO, PMDco and related ontologies | Materials, processes, properties, measurements, simulations and characterisation workflows |
| Measurements | QUDT and UCUM | Quantities, dimensions, unit systems, symbols and unit codes |
Agriculture, environment and marine science
AGROVOC
A multilingual FAO concept scheme supporting agricultural knowledge organisation and interoperability.
GEMET
A multilingual environmental thesaurus used for common environmental terminology in Europe.
NERC vocabularies
SKOS concept collections for oceanographic variables, instruments, methods and platforms.
Local domain extensions
Project-specific concepts linked to maintained environmental or disciplinary vocabularies.
Chemistry and life sciences
ChEBI
Identifiers and classifications for molecular entities, substances and chemical roles.
Gene Ontology
Structured terms for molecular functions, biological processes and cellular components.
BioPortal
A discovery and access environment for biomedical ontologies and ontology mappings.
Ontology annotations
Links experimental entities and metadata values to stable ontology concept identifiers.
Materials-science ontologies
| Resource | Role | Possible use |
|---|---|---|
| EMMO | Foundational ontology framework for applied sciences, materials modelling and characterisation | Align domain ontologies and represent materials, processes, measurements and physical properties. |
| PMD Core Ontology | Mid-level ontology for materials science and engineering | Describe objects, processes, measurements, workflows and digital materials data. |
| MatPortal | Repository and discovery environment for materials ontologies | Search existing ontologies before creating a local terminology. |
| OPTIMADE property definitions | Stable definitions of standard properties used by the OPTIMADE API | Align metadata describing computational materials structures and properties. |
Example: semantic annotation of a materials workflow
| Workflow element | Semantic annotation target |
|---|---|
| Material system | Material, chemical composition, crystalline structure, phase and defect concepts |
| DFT calculation | Simulation method, physical model, approximation, software and computational process |
| Training data | Dataset, structure, energy, force, stress and sampling concepts |
| MLIP model | Model, training process, model parameter, evaluation and software concepts |
| MD simulation | Molecular-dynamics process, ensemble, timestep, temperature and trajectory concepts |
| Provenance | Relationships between inputs, activities, agents, model versions and outputs |
Quantities and units
A quantitative metadata value should distinguish the numerical value, quantity kind, unit, uncertainty and measurement context.
Numerical value
The number recorded independently from its displayed unit and explanatory text.
Quantity kind
The measurable property, such as temperature, pressure, time, length or energy.
Unit
The standard measurement unit represented by a machine-readable code or URI.
Uncertainty
The uncertainty, tolerance, accuracy or confidence associated with the reported value.
QUDT and UCUM
| Aspect | QUDT | UCUM |
|---|---|---|
| Primary purpose | Semantic representation of quantities, units, dimensions and unit systems | Unambiguous coding of units for electronic data exchange |
| Representation | RDF vocabularies, classes, properties and resolvable URIs | Compact machine-readable unit codes |
| Quantity kind | Explicitly represented | Primarily inferred from the context in which the unit code is used |
| Dimensions and conversion | Can be represented semantically through structured properties | Defined through a formal unit-code system |
| Typical use | Knowledge graphs, linked data, ontologies and semantic metadata | APIs, laboratory systems, clinical data and machine-to-machine communication |
Example of a machine-readable quantity
{
"property": {
"label": "temperature",
"quantityKind": "http://qudt.org/vocab/quantitykind/Temperature"
},
"value": 300,
"unit": {
"label": "kelvin",
"symbol": "K",
"uri": "http://qudt.org/vocab/unit/K",
"ucumCode": "K"
},
"uncertainty": {
"value": 0.5,
"unit": "K"
}
}
Do not use: "temperature": "300 K" as the
only representation when structured metadata fields are available.
Prefer: separate fields for the quantity, numerical value, unit identifier and uncertainty.
How to record a domain concept
{
"value": "silicon carbide",
"conceptUri": "https://example.org/concept/silicon-carbide",
"vocabulary": "Example materials ontology",
"vocabularyVersion": "1.0",
"language": "en",
"mapping": [
{
"targetUri": "https://example.org/another-concept",
"mappingType": "closeMatch"
}
]
}Value: preferred label displayed to users.
Concept URI: stable identifier of the concept.
Vocabulary: authoritative semantic resource.
Version: release used when creating the metadata.
Mapping: relationship with a concept in another scheme.
Mapping between semantic resources
| SKOS mapping property | Meaning | Use with care |
|---|---|---|
skos:exactMatch |
The concepts can normally be used interchangeably across schemes. | Do not use merely because two labels are identical. |
skos:closeMatch |
The concepts are sufficiently similar for many applications but are not fully equivalent. | Preserve the distinction in high-precision scientific use. |
skos:broadMatch |
The mapped concept in the target scheme is broader. | Some domain specificity may be lost during export. |
skos:narrowMatch |
The mapped concept in the target scheme is narrower. | Verify that the narrower meaning is supported by the data. |
skos:relatedMatch |
The concepts are semantically associated but not equivalent. | It should not be treated as a substitution rule. |
Implementing a domain vocabulary
Define the field
Specify which scientific concept or quantity the metadata element represents.
Select the scheme
Identify the permitted vocabulary, ontology or unit system and its version.
Constrain the value
Define permitted branches, concepts, units, data types and cardinalities.
Validate the record
Check identifiers, vocabulary membership, units and semantic compatibility.
Recommended profile fields
| Field | Example content |
|---|---|
| Vocabulary name | Name of the thesaurus, ontology or unit vocabulary |
| Vocabulary URI | Persistent reference to the semantic resource |
| Vocabulary version | Release number or date |
| Concept URI | Persistent identifier of the selected concept |
| Preferred label | Human-readable term in a specified language |
| Local label | Interface label or translated label used locally |
| Mapping relation | Exact, close, broad, narrow or related match |
| Selection date | Date on which the concept was added to the profile |
| Concept status | Active, deprecated or replaced |
Validation of domain concepts and units
Identifier validation
Check the URI syntax, resolvability and expected concept type.
Scheme validation
Confirm that the concept belongs to the permitted vocabulary or ontology.
Scientific validation
Confirm that the term accurately represents the entity, method, property or process.
Unit validation
Check unit compatibility, dimensions, value ranges and required quantity kind.
Common implementation errors
Recording only a term label: the underlying concept cannot be identified reliably across languages and systems.
Selecting an ontology by name alone: its coverage, governance and implementation quality are not evaluated.
Using the broadest available term: scientifically important specificity is lost.
Using an overly specific term: the metadata asserts information that is not supported by the dataset.
Combining several concepts in one value: individual concepts cannot be indexed, mapped or validated.
Creating local synonyms as new concepts: duplicate concepts and conflicting identifiers are introduced.
Storing the unit inside a text string: automated validation and conversion become unreliable.
Ignoring deprecated concepts: metadata continues to use terms that the vocabulary maintainer has replaced.
Assuming identical labels mean identical concepts: semantic mappings become scientifically incorrect.
How to use this resource
1. Define the need
Identify the scientific entity, method, property, process or quantity that requires standardisation.
2. Find the resource
Search community standards, semantic registries and relevant domain repositories.
3. Record identifiers
Store concept URIs, labels, vocabulary names, versions and unit codes separately.
4. Review and maintain
Monitor new releases, deprecated concepts and changes in mappings.