Sweden Archaeology Site Discovery¶

The Sweden archaeology discovery surface answers a practical question: which SEAD sites can I inspect in Sweden, and what temporal and bibliographic support does this repository currently hold for each one? It publishes every geolocated Swedish SEAD site rather than only the best-dated subset.

The result is a discovery aid, not a league table of archaeological importance. Its order measures readiness for repository-based investigation. It says nothing about present excavation activity, preservation quality, research urgency, or the historical significance of a place.

Complete Population¶

Discovery population	Count
Swedish SEAD sites	1,925
sites with linked numeric chronology	370
sites with unresolved chronology	1,555
sites with linked bibliography	1,074
linked numeric map features	8,172
explicitly unresolved map features	1,555
total map features	9,727

One site can contribute several numeric features because each feature keeps a real grouped SEAD chronology interval. A site without eligible numeric chronology contributes one unresolved feature. Thus the site denominator is 1,925 while the map-feature denominator is 9,727. Sixty source-native post-1950 BP claims remain traceable but are refused by the canonical nonnegative-BP comparison contract.

flowchart TD
    Sweden["1,925 assigned Swedish SEAD sites"] --> Time{"linked numeric chronology?"}
    Time -->|yes: 370 sites| Intervals["8,172 interval-preserving features"]
    Time -->|no: 1,555 sites| Unresolved["1,555 explicitly unresolved features"]
    Intervals --> Atlas["9,727-feature Swedish discovery layer"]
    Unresolved --> Atlas

What “Discovery Readiness” Means¶

Every site enters one of four transparent tiers:

chronology and bibliography ready;
chronology ready;
bibliography ready; or
inventory only.

Within a tier, sites are ordered by chronology-kind breadth, represented chronology source-record count, bibliography count, dataset count, then stable name and SEAD identifier. The order helps a reader find sites for which the repository can already support a deeper evidence review.

It must not be relabelled “best sites,” “most important sites,” or “active sites.” A low-ranked site may be archaeologically exceptional but poorly represented by the captured relations. A high-ranked site may simply have a deep digital record.

Why RAÄ Does Not Supply Missing Sites Or Dates¶

The RAÄ artifact in this repository is a one-degree density surface. It can answer “how many published RAÄ records fall in this coarse cell?” It cannot identify which RAÄ records coincide with a SEAD point, establish a site's registry identity, or date that site.

Each discovery-registry row therefore carries the local RAÄ cell count only as raa_density_context_count. That field:

never changes discovery rank;
never creates a site;
never resolves chronology; and
never establishes current archaeological activity.

This refusal is intentional. Joining two source families by visual proximity would create an apparently concrete archaeology record whose identity and time neither source supplied.

How Time Works¶

A numeric feature keeps the exact time_start_bp, time_end_bp, midpoint, label, evidence class, uncertainty, and provenance locator from its normalized SEAD chronology feature. The discovery generator does not widen, average, or infer that interval.

An unresolved feature keeps null numeric fields and the comparability posture unresolved. At the full time extent it remains visible, so the map still supports spatial discovery. Under a narrowed BP window it cannot claim overlap. Its disappearance means “temporal eligibility unknown,” not “the site did not exist.”

stateDiagram-v2
    [*] --> FullExtent
    FullExtent: numeric and unresolved sites visible
    FullExtent --> NarrowWindow: select BP interval
    NarrowWindow: overlapping numeric intervals visible
    NarrowWindow: unresolved sites cannot assert overlap
    NarrowWindow --> FullExtent: restore full extent

Choose The Right Companion¶

Question	Artifact
inspect the complete ranking contract and registry	`data/sead/derived/sweden_archaeology_site_discovery.json`
sort or analyze one row per site	`data/sead/derived/sweden_archaeology_site_discovery.csv`
navigate actual intervals and unresolved sites on a map	`data/sead/derived/sweden_archaeology_site_discovery.geojson`
read a compact coverage and interpretation summary	`data/sead/derived/sweden_archaeology_site_discovery.md`
audit captured linked rows behind a site	`data/sead/raw/nordic_sites.json`

The four discovery companions are staged together in the Nordic Atlas bundle. The original normalized SEAD site and temporal layers remain governed source products, but the Atlas uses the discovery GeoJSON instead of loading those two overlapping layers alongside it.

A Safe Investigation Workflow¶

Start with the complete discovery layer and choose a place or time window.
Record the stable discovery feature ID and SEAD site ID from the popup.
Inspect the site's readiness tier, linked chronology kinds, bibliography count, and unresolved warnings.
Follow the SEAD source URL and raw captured relations before making a record-level claim.
Treat RAÄ density as regional context unless a separately governed record-level RAÄ acquisition supplies identities and chronology.
State whether the result concerns a site, a linked chronology interval, or the repository's current evidence readiness.

Claims The Surface Refuses¶

The discovery surface does not establish:

that a site is currently excavated, sampled, endangered, or actively researched;
that ranking position measures archaeological value;
that every record at a site belongs to each displayed interval;
that an unresolved site is absent from a selected period;
that RAÄ density represents past population or activity; or
that a nearby lake, pollen sequence, aDNA sample, SEAD site, and RAÄ cell share one historical cause.

Continue with the SEAD handbook for the observation-unit model, SEAD exports for source products, RAÄ exports for the density boundary, and maps for cross-layer interpretation.