Filter Variant Optimization Metrics and Station-Level Phenological Observations for Winter Wheat in Germany between 1993 and 2024

This dataset provides two complementary outputs of the PhenoPhaseR processing pipeline (Möller and Gerstmann, 2026) forwinter wheat (Triticum aestivum L., DWD Plant ID 202), covering all observed phenological phases (sowing to harvest) for the entire territory of Germany for the period 1993–2024. PhenoPhaseR is an open-source R workflow that couples the PHASE model (Gerstmann et al., 2016) - a growing-degree-day (GDD) approach with adaptive outlier filtering - to derive spatially consistent, quality-controlled point observations of phenological phase-start dates from the DWD volunteer observer network. The dataset components provided here correspond to processing Steps 5–6 (filter variant optimization) of the seven-step pipeline. What's new in this release Validator-clean RO-Crate metadata. Resolves the NovaCrate validation issues that affected earlier versions: DCAT 3 type-range warnings: dct:creator → dct:Agent, dct:publisher → dct:Agent, dct:license → dct:LicenseDocument, dct:spatial → dct:Location (each contextual entity now carries the required second @type). The prov:Activity type label on the #filtervariant-creation block has been removed because its schema is not resolvable in RO-Crate's context. The prov: properties (prov:used, prov:wasGeneratedBy, prov:startedAtTime, etc.) continue to work via the prov: namespace prefix in @context. Explicit @context mappings added for wasGeneratedBy and spatialResolution. Concept DOI used as the dataset identifier. The root identifier now points to the all-versions DOI 10.5281/zenodo.19483111 rather than a version-specific DOI, so citations remain stable across releases. Produced by PhenoPhaseR v1.3.0. Built using the publish-hook output of build_filtervariant_ro_crate(). The underlying shapefiles, OPT score CSVs, and diagnostic plots are byte-identical to the previous version — this is a metadata-only release. Note on file paths The ro-crate-metadata.json describes the crate using its intended directory structure (shapefiles/, opt_scores/, opt_scores/diagnostics/). Zenodo flattens directories on upload, so files are served at flat URLs (e.g. https://zenodo.org/records/<RECORD_ID>/files/DOY_202-10_2018.shp) rather than under the shapefiles/ subfolder. Tools that validate the crate against the original directory structure (e.g. NovaCrate) will report "file not found" warnings unless the data files are downloaded into matching subdirectories alongside the metadata file. This is a Zenodo display behaviour and does not affect the integrity of the data or its FAIR metadata.