AgricultureID

Methodology

Sources & methodology

This page explains where our information comes from, how sources are recorded and cited, and the automated checks that every entry must pass before it is published.

The source registry

AgricultureID maintains a structured registry of 74 sources. Each source records its organization, title, canonical URL, type, jurisdiction, language, access date, and a reliability classification. Content cites sources by a stable identifier, so the same authority can be referenced across many entries without duplication, and every citation resolves to a real registry entry.

Preferred sources

  • National agriculture ministries and government agencies
  • FAO and other recognised intergovernmental organizations
  • USDA and its agencies (ARS, NRCS, NASS, ERS, APHIS, FAS)
  • The European Commission and EFSA
  • Official plant-protection authorities such as EPPO
  • Universities, extension services, and agricultural research institutes
  • Peer-reviewed scientific literature
  • Official breed and crop registries

Sources we avoid

  • Anonymous blogs and unattributed content
  • Affiliate articles used as scientific sources
  • Content farms and copied material
  • Unsupported commercial claims
  • AI-generated or invented citations and URLs

How claims are handled

Because agricultural figures vary by region and season, we reference primary datasets rather than republishing volatile numbers, and we describe scale qualitatively where a precise figure would be misleading. Quantitative guidance that depends on local conditions — rates, dates, volumes, and yields — is framed as region-specific rather than universal.

Validation that gates publication

Content is authored as structured, typed data and checked automatically. Our validation fails the build if it finds any of the following:

  • A cited source that does not exist in the registry
  • Duplicate source identifiers or malformed source URLs
  • A published entry with no sources
  • Duplicate slugs, canonical URLs, titles, or descriptions
  • Related-content links that do not resolve
  • Missing required fields or malformed dates
  • Orphaned public entries or placeholder text
  • Structured-data that fails to generate

These checks run locally and in continuous integration, so structural problems are caught before anything is published. See our editorial policy for how we weigh evidence and treat uncertainty.