Evaluate and extend ISO field coverage in MDTranslator output

**What:** Based on SPIKE findings, implement mappings for ISO fields that have a viable DCAT-US v3.0 equivalent but are not currently output by MDTranslator. Category A fields (intentionally out of scope) are explicitly excluded.

**Likely candidates (to be confirmed by SPIKE-4):**

| ISO field | Likely DCAT-US v3.0 mapping |
|---|---|
| Topic category | `dct:subject` (skos:Concept with ISO topic category URI) |
| Feature catalog link | `conformsTo` (dct:Standard reference) |
| Lineage summary | `wasGeneratedBy` (Activity description, flattened) |
| Status / progressCode | No DCAT-US field -- document as intentionally dropped |
| Purpose | `description` (appended or separate field) |
| Scale denominator | `spatialResolutionInMeters` note (free text, not numeric) |
| Quality report text | `hasQualityMeasurement` (if lightweight mapping is feasible) |

Implementation approach depends on whether changes can be made to MDTranslator or whether they must be implemented as post-MDTranslator fixups in the harvester. MDTranslator is an external dependency; changes there require coordination with its maintainers and may be out of scope for the current contract.

Alternative: Implement additional fixups in `Record.harvest()` that extract specific fields from `self.source_raw` (the original ISO XML) after MDTranslator runs, and add them to `self.transformed_data` before validation and sync.

**Done When:**
- PM has identified which fields to pursue
- Chosen fields are mapped and populated in harvested records
- New fields pass v3.0 schema validation
- Field mapping decisions are documented in the harvester repo

**Notes:**
- Do not attempt to map CRS, spatial representation type, or positional accuracy -- these have no DCAT-US v3.0 equivalent and should be documented as intentionally dropped (Category A)
- GeoPlatform can access these fields from the original ISO record via the ISO-linkage distribution pattern; they do not need to be in the DCAT-US record