bluegreencity-ki

Citree Mapping Review

Purpose

This document records the reviewed mapping decisions for the Citree enrichment based on ga_lang.

Unlike enrichment.md, which describes the implemented runtime logic, this file explains why selected labels remain unmatched in the Hannover inventory snapshot.

Current Snapshot

The review below is based on the currently active inventory snapshot loaded by the pipeline (timestamp_filter: latest) and the current Citree export in data/derived/species/citree_species_criteria.csv.

Current status

Metric	Count
Unique ga_lang values	456
Matched to Citree	282
Unmatched (no_match)	172
Ambiguous (ambiguous, before review)	2

Ambiguous labels

Before this review, these labels matched more than one generic Citree candidate with the same score. They now remain unmatched because no cultivar-specific target exists.

ga_lang	Count	Decision
Ulmus-Hybride 'Clusius', Ulme 'Clusius'	11	Keep unmatched: the current export has no cultivar-specific target
Ulmus-Hybride 'Dodoens', Ulme 'Dodoens'	11	Keep unmatched: the current export has no cultivar-specific target

The current catalog contains generic and other cultivar rows for several Ulmus taxa, but no row for either reviewed cultivar. A deterministic tie-break would therefore select an unrelated profile rather than resolve duplicate representations of the same cultivar.

Review Categories

1. Non-species operational placeholders

These labels are not real taxa and should not be mapped to Citree profiles.

ga_lang example	Count	Decision
Objektkontrolle	1384	Treat as operational placeholder and keep unmatched; species_is_placeholder provides the explicit QA classification
schon ersetzt?	312	Keep unmatched; this is a workflow note, not a taxon

Additional labels of the same type should be reviewed with the same rule: if the value is a note, workflow marker, or missing-value surrogate, it must not receive a Citree profile.

2. Unknown-at-genus-level placeholders

These labels are taxonomically too unspecific for a safe Citree match.

ga_lang example	Count	Decision
Tilia spec., Linde (Art nicht bekannt)	643	Keep unmatched until a generic Tilia spec. profile exists or a domain-approved fallback is defined
Populus spec., Pappel (Art nicht bekannt)	345	Keep unmatched; no safe transfer to other poplar species
Quercus spec., Eiche (Art nicht bekannt)	326	Keep unmatched; species-specific tolerances differ too much
Acer spec., Ahorn (Art nicht bekannt)	306	Keep unmatched
Prunus spec., Kirsche oder Pflaume (Art nicht bekannt)	272	Keep unmatched

3. High-volume real taxa missing in Citree

These are real species or cultivar groups in the inventory, but the current Citree export does not expose a direct target profile for them.

Group	Current volume	Current issue	Decision
Populus nigra, Populus nigra 'Italica'	965	No direct Citree profile for Schwarzpappel / Säulenpappel	Keep unmatched; do not transfer from other Populus taxa
Populus canadensis and cultivars	225	No direct Citree profile for Kanadische Pappel	Keep unmatched; do not infer from hybrid or other poplar entries
Ulmus carpinifolia	123	Historical synonym of Ulmus minor, but no direct Feld-Ulme target in Citree	Keep unmatched until Citree exposes Ulmus minor / Feld-Ulme explicitly
Malus domestica	560	No safe Kulturapfel profile in the current Citree export	Keep unmatched; do not auto-map to ornamental crabapple profiles

Decision Rationale

Populus groups

Recommendation:

• Do not add a heuristic fallback from Populus nigra or Populus canadensis to any other Populus profile.
• If these species are needed for scoring, they require either:
  • an updated Citree export containing the exact taxon, or
  • a separate curated local mapping table agreed with domain experts.

Reasoning:

• Poplar taxa differ strongly in drought, heat, crown architecture, and site preference.
• A genus-level fallback would look technically successful but be fachlich weak.

Tilia spec.

Recommendation:

• Keep Tilia spec. unmatched.
• Only map it automatically if a domain decision explicitly approves a fallback order, for example to the statistically dominant lime species in the inventory.

Reasoning:

• Tilia cordata, Tilia platyphyllos, Tilia tomentosa, and hybrids can differ materially in tolerance values.
• A generic lime profile does not currently exist in the Citree export.

Ulmus carpinifolia

Recommendation:

• Keep unmatched for now.
• Track it as a taxonomic synonym candidate (Ulmus carpinifolia → Ulmus minor) but do not map further until a direct Citree profile exists.

Reasoning:

• The synonym itself is well known, but the target species still has no direct Citree row in the current export.
• Mapping to other elm hybrids would overstate certainty.

Malus domestica

Recommendation:

• Keep unmatched unless the team explicitly decides that ornamental crabapple profiles are an acceptable proxy for Kulturapfel.
• If such a decision is taken, record it explicitly in enrichment.md with justification.

Reasoning:

• Edible apple cultivars and ornamental Malus forms should not be collapsed silently.
• The current export does not provide a clear, direct target for Malus domestica.

Ambiguous Ulmus cultivars

Decision:

• Keep Clusius and Dodoens unmatched because the current Citree export has no corresponding cultivar rows.
• Do not add a genus-level tie-break or alias to another elm cultivar.

Reasoning:

• These labels are low-volume compared with the poplar and placeholder groups, but they are precise cultivar labels.
• They are good candidates for a future targeted rule once the correct Citree target is verified.

Already Resolved In Code

The following formerly problematic taxa are now handled explicitly by the runtime matcher and do not require further review at this time.

Inventory form	Runtime handling
Platanus x acerifolia	Alias to Citree Platanus x hispanica
Ulmus hollandica	Alias to Citree Ulmus x hollandica
Quercus petraea	Alias to the older Citree taxonomy under Quercus robur subsp. petraea
Liquidambar styraciflua	Alias to the Citree spelling variant Liquidambar styraciflura

Review Outcome

• Operational placeholders remain no_match and are identified through species_is_placeholder.
• Unknown-at-genus-level labels remain no_match.
• Populus nigra, Populus canadensis, Tilia spec., Ulmus carpinifolia, and Malus domestica remain unmatched by design.
• Clusius and Dodoens remain no_match until exact cultivar rows become available.
• No proxy mapping or local supplement is approved by this review.

Maintenance Note

Whenever the Citree export is refreshed, this file should be revisited together with the citree_match_rule distribution from a fresh pipeline run. If one of the groups above becomes directly available in Citree, the runtime matcher should prefer an explicit botanical match over any proxy logic.