An environmental science concept defining interactions among organisms, resources, and human pressures. It governs how disturbances and management actions change ecosystem structure, function, and service delivery. It does not ensure desired outcomes without monitoring, enforcement, and adaptive management of pressures. It supports protection and restoration by translating ecological evidence into measurable objectives and actions. The concept is generally stable, though metrics and monitoring methods improve over time.

Principle

Adhere to data stewardship best practices (FAIR: findable, accessible, interoperable, reusable), include provenance and quality information, document sampling and processing methods, and provide machine-readable formats and controlled vocabularies where possible.

Demonstration

A publicly released dataset containing hourly NO2 concentrations from 50 monitoring stations for 2010–2020 in CSV and NetCDF formats, with station coordinates, instrument types, calibration records and QC flags.

Misapplication

Publishing raw instrument counts or aggregated averages without metadata, calibration factors or quality indicators, or merging heterogeneous datasets without harmonization and provenance, leading to incorrect analyses.

Consequence

Well-constructed datasets enable reproducible research, model calibration and validation, trend analysis, meta-analyses and transparent policy evaluation; poor datasets create false confidence and misinformed decisions.

Reversal

Ad-hoc, undocumented observations or closed proprietary records prevent reuse and scrutiny; conversely, excessive restriction on data access can limit validation and application.

Boundary

Covers observational records, processed measurement series and documented model outputs intended for analysis; excludes ephemeral unrecorded observations, undocumented proprietary data lacking metadata, and raw instrument data without context unless accompanied by processing documentation.

Semantic Tension

Tension between dataset and database/stream: a dataset is a curated, versioned snapshot intended for analysis, while a database may be dynamic and a stream supplies continuous measurements; careful distinction is needed when using for modeling or regulatory purposes.

Synthesis

A pollution dataset is an organized, documented and quality-controlled collection of pollution-related data that makes empirical information discoverable, interpretable and reusable for analysis, modeling and decision support.