Skip to content

Neotoma LRA prerequisite review

The source-bound projection has SHA-256 26135d91bbe5d729dcca3a1f6f0ea69f5cec4dc00a54a6dd740d5004d4e4b809. It retains 10,637 sample-owned chronology review rows and 163 source-site basin review rows from the corrected Neotoma relational snapshot. Country partitions are Sweden 5,011 samples/77 sites, Denmark 91/1, Norway 3,544/52, Finland 1,797/28, and unassigned 194/5. The unassigned partition is not a fifth Nordic country. Every row carries source identity, review status and source locators.

The lakeparameter source capture contains 24,554 rows in verified public API pages. It was originally bound to a 54-site review with SHA-256 aaded8ed751a8f245a357c50bb2c2a96a8710168deab7223f0c48284180e96ce. The producer verifies that original review and the original capture summary against the exact page bytes. It then verifies that all 54 original site IDs remain in the corrected 163-site cohort under the same raw Neotoma snapshot, and projects the same source pages against the larger cohort. Fourteen native lakeparameter rows match 13 current sites; 43 sites separately retain a native site area value without an established unit or historical basin meaning. The source pages came from live offset pagination, not a transactional snapshot; absence of a row is not certified.

These are review leads. Chronology choice, taxon classification, count support, PPE/fall-speed binding, justified basin assumptions and qualified returned decisions remain unresolved. The projection has zero accepted samples, scientific_acceptance=not_evaluated, and no executable LRA request. It does not supersede the separate four-country observed-pollen prediction study.

Historical correspondence and a new reconstruction

Reproducing LandClim requires its actual selected chronology, sample membership, geometry and parameters. A separately declared reconstruction can use a source-qualified alternative chronology and a justified static basin model. Trondman et al. (2016) assumed constant present basin size and type where stratigraphic information was insufficient, with exceptions where evidence supported a change. LandClim II describes several chronology sources and circular radius approximations. These published choices support explicit modeling assumptions; they do not authenticate a particular site's past shoreline or the historical LandClim selection.

A new profile must name its question, physical core and source rows, permitted use, chronology selection, taxon mask, count support, parameter units, geometry, pooling and weighting, sensitivity and evaluation target. Missing historical selection blocks historical correspondence, while a justified alternative can answer its separately stated question. An interval crossing a time-bin boundary can contribute to conditional set-based bounds without an invented probability distribution. Such bounds are distinct from a selected age model and joint confidence intervals. A common denominator cancels from a fixed-mask normalized REVEALS point estimate; original counts and taxon eligibility still govern sampling uncertainty and the reconstructed composition. None of these distinctions admits the presently unreviewed source rows automatically.

To reproduce, run python -m bijux_pollenomics_dev.taxa_review.lra_prerequisites with the repository raw root data/neotoma/raw/neotoma_pollen_dataset_downloads, normalized root data/neotoma/relational, external capture root <external-Neotoma-lakeparameter-capture-root>, capture SHA-256 0c800ed252d6bd2bbed72f5ac7798e27d7a72f89330170b8ce33989384c758fd, and a new output path under artifacts/. Compare the resulting exact SHA-256 with the governed projection above before reuse. The external capture path is a local custody location, not a source publication URL.