Skip to content

SEAD

SEAD supplies environmental-archaeology context. In the current checked-in state it is a spatial site inventory, not a uniformly time-resolved archaeology layer. Its value is substantial precisely when that contextual role and temporal limit remain explicit.

Checked-In Evidence

The current temporal and legibility reviews cover 2,195 captured site rows:

Property Count Posture
site inventory rows 2,195 spatial archaeology context
numeric temporal intervals 0 no numeric same-period comparison
linked dating-range rows 0 dating foundation not captured here
linked relative-period rows 0 period foundation not captured here
bibliography rows 0 bibliography linkage not captured here

The normalized spatial layer contains 2,172 records. The difference between captured and normalized counts must remain visible; it is not evidence that the omitted records never existed.

Every current legibility-review row is classified as inventory-only or unresolved for time, with high risk from the thin site-inventory capture and a publication posture of context with an explicit caveat.

Curation Lineage

Boundary Governing material Decision preserved
inventory capture raw/nordic_sites.json the site identities and geometry available in the checked-in snapshot
spatial normalization normalized/nordic_environmental_sites.geojson the 2,172 members eligible for contextual mapping
access review review/access_model.json how the captured surface relates to deeper SEAD information
temporal review review/temporal_review.json refusal of unsupported numeric comparison
evidence legibility review/evidence_legibility_review.json why inventory visibility is not record-level interpretive depth
recovery requirements review/recovery_requirements.json which missing relations would materially strengthen future claims and how their recovery is demonstrated

These surfaces turn a thin database capture into an auditable evidence state. They do not conceal the missing chronology and bibliography behind a successful normalization count.

Database Preparation Is Selective Materialization

SEAD is relational upstream, while the current governed capture materializes a site inventory. The preparation contract therefore records not only the rows that exist, but the relations that were not captured. That distinction prevents an inventory-shaped export from being mistaken for the full source database.

flowchart LR
    Site["site inventory row"] --> Identity["stable site identity"]
    Site -. not captured .-> Dating["dating and relative-period relations"]
    Site -. not captured .-> References["bibliography relations"]
    Site -. not captured .-> Proxies["dataset and analysis relations"]
    Identity --> Point["normalized contextual point"]
    Dating --> Time["time-aware eligibility review"]
    References --> Attribution["record-level source interpretation"]
    Proxies --> Evidence["proxy-specific environmental claim"]

Only the solid path is currently governed for all captured members. The dashed relations describe evidence requirements, not implied empty tables. Recovering one relation strengthens only the claim dimension it owns: a bibliography link does not provide chronology, and a date does not establish which proxy or specimen produced it.

flowchart LR
    Capture["captured site inventory"] --> Normalize["normalized site points"]
    Normalize --> Temporal["temporal review"]
    Normalize --> Legibility["evidence legibility review"]
    Temporal --> Context["context-only publication posture"]
    Legibility --> Context
    Context --> Product["archaeology context layer"]

What SEAD Supports

  • environmental-archaeology context around samples, pollen records, lakes, and regions;
  • spatial density and proximity comparisons under declared distance bands;
  • cross-regional context beyond one national registry;
  • identification of places where deeper source recovery may be valuable;
  • landscape interpretation that keeps archaeology visible beside biological and environmental evidence.

What The Current Capture Does Not Support

  • same-period claims between SEAD sites and nearby pollen or aDNA;
  • exact sample identity, locality, chronology, or species assignment;
  • duration or phase comparison across the captured inventory;
  • bibliography-backed interpretation for every normalized site;
  • treating site density as a direct measure of past activity or preservation.

The absence of numeric intervals is not repaired with inferred dates. SEAD can affect spatial decision support while receiving no chronology credit.

Interpret The Capture Gap

The difference between inventory visibility and evidentiary depth is itself a governed result:

State What is known What remains unavailable
captured inventory row a source site identity entered the repository linked dating, period, bibliography, and deeper record relations
normalized site point a spatial representation passed family normalization uniform numeric chronology and record-level interpretive depth
contextual publication member the point is eligible for declared archaeology context same-period support or direct association with nearby evidence
omitted captured row the upstream inventory included a row not present in the normalized point layer the exact normalization, identity, or geometry issue must be inspected 
flowchart LR
    Inventory["2,195 captured rows"] --> Normalize["2,172 normalized points"]
    Normalize --> Context["spatial archaeology context"]
    Inventory --> Gap["23-row normalization difference"]
    Context --> Recovery["deeper chronology and bibliography recovery"]
    Gap --> Recovery

The 23-row difference is not automatically an error or a valid deletion; it is a review boundary that requires member-level explanation. Likewise, the lack of linked chronology is a source-recovery limitation, not permission to assign dates from nearby records or broad archaeological expectations.

Account For The 23-Row Difference

The current normalized layer requires a coordinate to classify inside one of the four publication-country boundary geometries. All 23 captured rows absent from the normalized layer retain coordinates but do not satisfy that country membership predicate. The set includes records east or south of the four country scope, such as Troitski, Hedeby, Riga, and Stralsund, as well as island or edge cases such as Tjärnan, Syllöda, Otterböte, and Kastelholm.

flowchart LR
    Capture["2,195 captured site identities"] --> Coordinates{"usable coordinates?"}
    Coordinates -->|yes| Country{"inside one publication-country geometry?"}
    Country -->|yes| Normalized["2,172 normalized Nordic points"]
    Country -->|no| Outside["23 captured, non-member rows"]
    Outside --> Audit["identity and coordinates retained for audit"]

The difference is therefore a scope-membership result in this snapshot, not deduplication and not evidence deletion. A future boundary or country-scope change must reevaluate those 23 identities; it must not silently append them or rewrite their coordinates. Report both captured and normalized totals whenever the distinction affects a coverage claim.

Read Absence At The Correct Boundary

SEAD exposes several forms of “not present,” and they are not interchangeable. The boundary that produced the absence must travel with the interpretation.

Observation What it establishes What it does not establish
no dating rows in the governed materialization this snapshot cannot support a linked numeric chronology claim the upstream SEAD database or the archaeological site has no dating evidence
no bibliography rows in the governed materialization this snapshot cannot provide record-level bibliography lineage no publication discusses the site
captured row absent from the normalized layer the row did not meet the current four-country spatial membership rule the source identity is invalid or was deduplicated
normalized point absent from a selected map view the current product or interaction did not display it the point is absent from the normalized family population
no nearby SEAD point under a declared radius no admitted member met that spatial query no archaeological evidence exists near the target place 
flowchart TD
    Missing["apparently absent evidence"] --> Boundary{"where is it absent?"}
    Boundary -->|capture| Materialization["not materialized in this snapshot"]
    Boundary -->|normalization| Membership["outside current membership rule"]
    Boundary -->|publication| Product["not admitted to this product"]
    Boundary -->|interaction| View["not visible in current view"]
    Materialization --> Claim["boundary-qualified absence claim"]
    Membership --> Claim
    Product --> Claim
    View --> Claim

This distinction is central to database curation. A governed thin capture is useful because it makes its selection boundary inspectable; it must not be described as though it exhaustively represents every relation in the upstream system.

Relationship To RAÄ

SEAD provides wider environmental-archaeology context. RAÄ provides denser Sweden-specific registry context. Their records may overlap spatially, but the families have different coverage, source systems, and normalization semantics. They should be compared as complementary context rather than merged into a single archaeology truth set.

Choose SEAD For The Question

Question Use Retain with the claim
Which environmental-archaeology sites are spatially near this feature? normalized site points under a declared distance rule SEAD member identity, distance, and inventory-only posture
Were nearby records contemporaneous? not from the current capture linked dating or period evidence must first be recovered and reviewed
Does a dense cluster represent greater past activity? not directly capture, investigation, preservation, and database-selection effects remain alternatives
Where would deeper source recovery add the most value? inventory, gap, and recovery-review surfaces together the missing relation and the claim it would unlock

Review A Refresh Against The Access Boundary

SEAD refresh review first distinguishes a changed upstream inventory from a changed accessible projection. It then accounts for raw identities, normalization membership, the 23-row boundary, access-model changes, temporal eligibility, and published descendants. A member becoming accessible is not evidence that its missing chronology was recovered; access and scientific fitness are independent transitions with separate receipts.

Governing Surfaces

  • data/sead/raw/nordic_sites.json records the captured inventory;
  • data/sead/normalized/nordic_environmental_sites.geojson governs normalized points;
  • data/sead/review/access_model.json records access posture;
  • data/sead/review/evidence_legibility_review.json records interpretability;
  • data/sead/review/temporal_review.json records temporal refusal;
  • data/sead/review/recovery_requirements.json records the evidence gaps, required evidence, and satisfaction signals for stronger use.

Together these files form a database-preparation receipt: captured population, normalized membership, access posture, temporal eligibility, interpretation risk, and the evidence required for stronger use remain independently inspectable.

The SEAD handbook expands the interpretation and collaboration context without changing these evidence limits.