Enriching Weather Index Metadata for Agricultural Decision Support: Standardized Quality, Uncertainty, and Data-Fitness-for-Purpose Assessment (DFFP)

Agricultural decision support - from weather index insurance to yield-gap analysis - depends on geodata whose quality and fitness for purpose can be assessed transparently. In practice, however, quality information is scattered across documentation, pixel-level uncertainty remains unquantified, and fitness-for-purpose reasoning is left to individual users without structured guidance. We demonstrate a workflow that addresses all three gaps by enriching weather index metadata as machine-actionable FAIR Digital Objects. Two independent uncertainty sources - the PHASE interpolation error and the E-OBS ensemble spread - are propagated through the processing chain into per-pixel 1σ bands co-delivered with every annual product. Three metadata components are embedded into RO-Crate 1.2 containers published on Zenodo: (1) ISO 19157-1 accuracy metrics per crop, phase, and year; (2) spatial uncertainty layers enabling reliability assessment against user-defined thresholds; and (3) a fitness-for-purpose application matrix, LLM-extracted from published studies, documenting validated use contexts and limitations. Together, these components enable users to ask: "Where are these data sufficiently accurate and validated for my specific decision context?"