Swedish pollen cohort availability

We checked the Neotoma Swedish pollen dataset query captured on 1 October 2026 against the retained relational site identities. Its 102 dataset records identify 99 distinct sites. All 99 site IDs are already in the retained 200-site snapshot. This establishes no new physical site within that query; it does not rule out another archive or a future deposit.

We also checked two PANGAEA pollen metadata leads. The Sämbosjön SAMBOTOT pollen record shares its core handle with Neotoma site 3398, collection unit 4262 and pollen dataset 4403, which has already been scored. The Klotjärnen KLOT pollen record shares its core handle with retained Neotoma site 13336, collection unit 14908 and pollen dataset 19909. Their archive coordinates differ, so we preserve both coordinates in the machine result; the handle and linked site metadata are enough to refuse either as an independently fresh core, not to settle the coordinate discrepancy. The SAMBOTOT dated-control record lists depths and raw radiocarbon ages/errors. Its checked metadata does not supply a selected calibrated pollen-sample age model.

We then screened a bounded PANGAEA Swedish pollen-profile metadata search. Of 66 returned titles, 62 named handles match retained Neotoma collection units, three named handles do not, and one generic Schonen title has no bound core handle. We compared all 206 retained core handles, including records whose governed country assignment is UNASSIGNED. Restricting that comparison to SE alone would incorrectly present three existing cores as new. Source geopolitical labels still do not approve a country assignment.

Handle	Pollen DOI	Retained Neotoma site/core/dataset	Age-determination lead	Access declaration
ADRAN	738941	No exact handle match	740137	Access rights needed
GARASEL	739248	No exact handle match	740365	Access rights needed
HAKULLA5	711892	3138 / 3955 / 4082; UNASSIGNED	740389	CC BY-NC 3.0
LEVEANIE	739449	No exact handle match	740509	Access rights needed
NDRAKET	739584	3311 / 4162 / 4298; UNASSIGNED	No age-determination title in checked handle search	CC BY-NC 3.0
VUOLEPAL	889510	26118 / 32167 / 45312; UNASSIGNED	888917	Access rights needed

The Giesecke final sample-age table lists 37 Swedish event handles; all 37 match retained Neotoma handles, and none is one of these six. The Fyfe chronology-reference table identifies HAKULLA5 as using “EPD chron2” and cautions that its top may be less accurate. The later LegacyAge site table independently identifies HAKULLA5 as Neotoma site 3138/dataset 4082 with its own revised AWI age model. Neither site metadata table is a selected pollen-sample age series for a new core. The three unmatched handles are only leads, not proven distinct physical sites. GARASEL's age-determination lead, the EPD data access route, or a source-owner packet would need to establish core identity, selected sample ages and permitted use before any prediction freeze.

We checked GARASEL's linked records more closely. Its pollen-profile metadata identify depths from 0.12 to 5.50 m and pollen count parameters, but no sample-age column. The dated-control metadata identify depths from 3.025 to 7.175 m and raw radiocarbon dates and errors, not a selected calibrated age at each pollen-sample depth. Both records declare unknown licensing and require login for data access. A common event name is a lead for core reconciliation, not proof of an independent physical core. GARASEL still contributes zero admitted prediction samples.

We opened metadata only, before any new candidate pollen observation or count table. This result does not validate an independent holdout or admit LRA inputs. The next candidate needs a distinct physical core, a source-linked selected calibrated sample-age sequence, permitted use, and a frozen prediction protocol before target counts are inspected. The research need remains open.

The source declaration binds seven Neotoma/PANGAEA responses, 17 EPD/PANGAEA metadata and age-table sources, and the relational snapshot by exact digest. Reproduce the machine result from the repository root, with the private source store adjacent to the repository:

PYTHONDONTWRITEBYTECODE=1 .venv/bin/python \
  docs/report/source-relations/neotoma/swedish-cohort-availability/replay.py \
  --repository . --private-root ../bijux-pollenomics-data \
  --output artifacts/jobs/swedish-cohort-availability/replay.json
cmp docs/report/source-relations/neotoma/swedish-cohort-availability/report.json \
  artifacts/jobs/swedish-cohort-availability/replay.json