Sweden Lake Priorities¶
The Nordic atlas now carries optional Sweden lake ranking overlays so readers can inspect the strongest lake candidates without leaving the shared map. These layers stay off by default because they are a ranked interpretation surface, not a base evidence surface. The governing packet still lives in the Sweden report tree, and the atlas overlay should only help readers inspect the same ranked packet in geographic context.
Use this page when you need to understand what the Sweden lake overlays are actually ranking, what they are not ranking, and how far those scores should be trusted before field planning.
What The Overlay Shows¶
The public Sweden lake packet currently ranks 6,763 registry-backed lake candidates. The Nordic atlas exposes the following optional overlays:
- aggregate top 40
- consensus top 40
- 10 km top 40
- 20 km top 40
- 30 km top 40
- 40 km top 40
- 50 km top 40
- fieldwork shortlist top 20
The fieldwork shortlist remains top 20 because that is the governed public shortlist published by the Sweden packet today. The other scenario layers now expose the top 40 rows directly on the map.
How Ranking Works¶
The candidate pool comes from the Sweden lake registry published through SMHI SVAR. Each lake uses one representative point derived from the official lake polygon rather than a pollen-site centroid. Only lakes with at least one human aDNA locality within 50 km remain in the ranked set.
Aggregate ranking blends the five radius scenarios with these weights:
| Radius scenario | Weight |
|---|---|
| 10 km | 0.35 |
| 20 km | 0.27 |
| 30 km | 0.18 |
| 40 km | 0.12 |
| 50 km | 0.08 |
Inside each scenario, the public packet weights evidence families like this:
| Evidence component | Weight |
|---|---|
| Human aDNA signal | 0.59 |
| Direct pollen signal | 0.14 |
| Nearby pollen signal | 0.07 |
| Lake sampling fit | 0.07 |
| Archaeology signal | 0.07 |
| Domesticated animal signal | 0.04 |
| Evidence diversity signal | 0.02 |
The decision rule is intentionally conservative:
- human aDNA locality and sample coverage decide the ranking first
- direct pollen support is the next tie-break
- broader pollen and archaeology context come after that
- sampling fit and blended score resolve later ties
The consensus overlay is different from the aggregate overlay. It lifts lakes that recur across multiple scenario top slices, then breaks ties by the mean scenario rank and finally by aggregate rank.
What These Scores Do Not Include¶
The public ranking is useful, but it is deliberately narrower than a full site selection workflow. It does not currently ship:
- governed bathymetry or coring-depth surfaces
- shoreline access, permits, or landowner logistics
- a field-confirmed judgment that one lake is already ready for sampling
- a claim that the highest-scoring row is the best scientific target under all practical constraints
What Lake Properties Are Public¶
Each mapped lake row currently ships these reader-facing properties:
- registry-backed coordinates
- aggregate rank and aggregate score
- scenario rank and scenario score
- lake registry identifier
- lake area in square kilometers
- sampling posture and sampling fit
- duplicate-name and coordinate-spread diagnostics
- supporting scenario-presence counts
This is enough to compare evidence density against basic lake suitability, but
it is not a full limnology packet. The current public packet does not
ship governed depth, width, bathymetry, coring logistics, or access-permit
surfaces. Where a row says compact_lake_candidate, small_lake_review, or
sampling_lake_candidate, that is a ranking posture, not a substitute for
field bathymetry or basin-shape verification.
Current High-Ranking Lakes¶
The public atlas page should let readers see concrete examples quickly, so the table below mirrors the top of the current aggregate registry packet.
| Aggregate rank | Lake | Aggregate score | Scenario top-20 presence | Area km² | Sampling posture |
|---|---|---|---|---|---|
| 1 | Bergsjön | 0.5947 | 4 | 0.063346 | compact_lake_candidate |
| 2 | Hulesjön | 0.5875 | 3 | 0.037617 | small_lake_review |
| 3 | Sjötorpasjön | 0.5862 | 7 | 0.603122 | sampling_lake_candidate |
| 4 | Hornborgasjön | 0.5037 | 6 | 27.925549 | sampling_lake_candidate |
| 5 | Skårsjön | 0.4818 | 3 | 0.021492 | small_lake_review |
| 6 | Rösjön | 0.4651 | 6 | 0.956929 | sampling_lake_candidate |
| 7 | Bjärsjön | 0.4573 | 6 | 0.132579 | compact_lake_candidate |
| 8 | Tresjö | 0.4433 | 3 | 0.104225 | compact_lake_candidate |
| 9 | Vartoftasjön | 0.4186 | 4 | 0.154654 | sampling_lake_candidate |
| 10 | Alasjön | 0.4053 | 4 | 0.191045 | sampling_lake_candidate |
| 11 | Ullstorpasjön | 0.3983 | 4 | 0.290248 | sampling_lake_candidate |
| 12 | Häckebergasjön | 0.3859 | 3 | 0.758596 | sampling_lake_candidate |
Readers who need the full ranked set should use the CSV and markdown artifacts below instead of treating this table as the entire packet.
How To Read The Layers¶
- Use the aggregate layer when you want the strongest blended ranking.
- Use the consensus layer when you want lakes that stay near the top across multiple ranking scenarios.
- Use the radius layers when you want to inspect how nearby evidence density changes as the search window expands.
- Use the fieldwork shortlist when you want the current public set that already leans toward sampling practicality and reduced identity ambiguity.
When a lake looks promising on the map, the next step should be the report packet, not a stronger claim. The atlas helps you inspect the ranking in geographic context; the governing evidence still lives in the Sweden report artifacts linked below.