SMHI SVAR¶

SMHI SVAR supplies official Swedish lake identity and geometry for sampling-oriented products. It answers which registered water body is the candidate? It does not answer whether that lake contains suitable sediment, can be accessed, can be permitted, or is scientifically preferable.

SVAR is a sampling-domain family rather than direct biological or historical evidence. Nearby pollen, ancient DNA, and archaeology remain owned by their source families even when they contribute to a lake ranking.

Current Capture¶

Property	Governed value
source	SMHI SVAR
source surface	`https://vattenwebb.smhi.se/svarwebb/`
acquisition interface	WFS `lakes` type
capture date	`2026-06-22`
authority status	refused
admitted lake denominator	unavailable
source-reported count	retained in the capture receipt, not admitted as a governed denominator
publication layer key	withheld until authority admission
temporal posture	no time dimension

SVAR authority is refused because the governing normalized registry is absent. This unavailable governed denominator is not a claim of zero lakes or zero scientific relevance. It prevents source-reported summary counts from being promoted into current analytical or publication coverage.

Shipped And Declared Surfaces¶

The source-family contract declares a normalized registry at data/svar/normalized/sweden_lake_registry.geojson. The current repository snapshot does not ship that complete registry file. It ships:

data/svar/raw/svar_lake_registry_manifest.json, which records source, interface, acquisition date, and matched and normalized counts; and
data/svar/normalized/svar_summary.json, which retains a source-reported count and svar-lakes layer identity without establishing current authority; and
data/svar/review/sweden_lake_candidate_registry.geojson, a compact evidence-linked review registry with 99 unique official lakes, mapped areas, stable identities, and sampling-readiness gaps.

The compact review registry contains 96 evidence-linked official lakes plus the official matches for four named southern Sweden targets; overlap between those sets produces 99 unique members. It is deliberately a review surface, not a misleading partial replacement for the complete normalized registry.

Published Sweden ranking tables retain member-level SVAR identifiers, representative coordinates, source URLs, name diagnostics, water identities, and mapped areas for the candidates they publish. Those product members are auditable, but they are not a substitute for a portable checked-in copy of the complete normalized registry.

flowchart LR
    WFS["SMHI SVAR WFS"] --> Capture["capture manifest<br/>source-reported count"]
    WFS --> Registry["declared normalized registry"]
    Capture --> Summary["checked-in count summary"]
    Registry --> Authority["authority admission"]
    Authority --> Review["99-lake evidence-linked review registry"]
    Review --> Candidate["stable lake candidate identity"]
    Candidate --> Ranking["Sweden ranking and sensitivity products"]
    Ranking --> FieldReview["identity, basin, access, permit, and field review"]

The trust boundary is explicit: the manifest and summary preserve a historical source-reported count, while the compact registry establishes only its 99 reviewed members. Neither establishes the absent complete normalized registry or authorizes a source-wide SVAR publication denominator.

Lake Identity Contract¶

Field	What it establishes	What it does not establish
registry ID or UUID	stable source identity for the water body	unique human-readable name
water identity	source-owned water object	sampling suitability
official geometry	present registry extent	historical shoreline or depositional basin
representative point	reproducible anchor derived from the polygon	optimal access or coring coordinate
mapped area	geometry-derived screening input	depth, volume, sediment thickness, or preservation
source URL	route back to the registry object	permanence of every displayed attribute

Duplicate lake names are expected and must remain visible. A clean name match cannot replace the registry identifier, UUID, geometry, and coordinate when a candidate is cited or reviewed.

Lake, Candidate, And Sampling Location Are Different Objects¶

Object	Governing identity	Supported claim
registered lake	SVAR registry ID, UUID, water identity, and source geometry	which official water body the record represents
ranking candidate	product version, lake identity, representative point, mapped area, scenario, and evidence packet	how the lake compares under one declared decision-support model
fieldwork target	reviewed access, basin, bathymetry, sediment, permit, safety, and sampling rationale	where a field team may investigate
sampling station	field-owned coordinate, method, date, and observation record	where a specific visit or sample was actually made

flowchart LR
    Lake["registered lake identity"] --> Candidate["ranked candidate"]
    Candidate --> Review["field feasibility review"]
    Review --> Target["fieldwork target"]
    Target --> Visit["sampling station or visit"]

The transitions are decisions, not aliases. The representative point used for distance calculations must not be copied forward as a coring coordinate, and a high-ranked lake does not become an approved field target without the missing field-owned evidence.

From Registry To Decision Support¶

flowchart LR
    Lake["SVAR lake identity and geometry"] --> Eligibility{"candidate eligibility"}
    Eligibility -->|excluded identity class| Outside["outside shortlist"]
    Eligibility -->|human aDNA within 50 km| Bands["10–50 km evidence bands"]
    Bands --> Aggregate["aggregate ordering"]
    Bands --> Consensus["scenario consistency"]
    Aggregate --> Preparation["fieldwork-preparation screen"]
    Consensus --> Preparation
    Preparation --> Decision{"expert field decision"}

The registry owns the candidate, not the surrounding evidence. Distance bands join governed records around the candidate under a declared model. The model may prioritize review; it cannot convert registry presence into a sampling recommendation.

Audit A Lake Candidate¶

Record the publication version, ranking surface, and scenario.
Resolve the row to its SVAR registry ID, UUID, water identity, and source URL; use coordinates to disambiguate repeated names.
Confirm the representative-point method and mapped area.
Read the nearby evidence as separately owned source-family inputs.
Inspect aggregate, consensus, sensitivity, and preparation outputs rather than quoting an ordinal rank alone.
Treat bathymetry, sediment preservation, access, permits, and field safety as independent evidence required before a sampling decision.

Review A Refresh From Registry Identity Outward¶

Start with SVAR IDs, UUIDs, source URLs, names, and water geometries. Then compare the representative-point method and only afterward recompute eligibility, distance bands, rankings, and fieldwork-preparation surfaces. A changed centroid can alter every proximity result without changing the lake identity. A changed rank can leave the registry record untouched. The refresh receipt names which boundary changed instead of attributing every difference to “the lake.”

Governing Surfaces¶

data/svar/raw/svar_lake_registry_manifest.json governs capture identity and source-scale counts;
data/svar/normalized/svar_summary.json governs the checked-in summary;
data/svar/review/sweden_lake_candidate_registry.geojson governs the evidence-linked candidate population, mapped areas, and sampling gaps;
data/source_family_contracts.json declares lifecycle ownership;
data/source_spatiotemporal_posture_registry.json declares sampling-domain, distance, and no-time postures; and
Sweden lake priorities explains the derived ranking, sensitivity, and field-review boundaries.

Continue with source comparison before combining SVAR with another family and publication reports when reusing a ranked candidate.