Sweden lake archaeology-weight sensitivity

This report isolates how much the lake shortlist changes when archaeology is treated conservatively, at the published baseline, or with added contextual emphasis. It does not turn nearby archaeology into proof of lake suitability.

Why The Baseline Is 0.07

Archaeology receives 0.07 because it is contextual rather than direct lake or sampling evidence. It equals nearby-pollen and area-screen weight, remains below direct pollen, and remains far below human aDNA.

The archaeology component is itself bounded: SEAD site density contributes 0.40 inside the archaeology signal, numeric SEAD chronology 0.20, SEAD chronology overlapping nearby human aDNA 0.20, and coarse RAÄ density 0.20. When archaeology weight changes, every non-archaeology component is scaled proportionally so profile weights continue to sum to 1.00.

Decision Rule

A lake should move upward on archaeology grounds only when linked SEAD records are inspected and temporally compatible. RAÄ density alone is a discovery prompt, never sufficient promotion evidence.

Each sensitivity row publishes the candidate's available temporal context windows and the count of nearby numeric SEAD records. A weight change remains a static robustness check; interpretation in a particular period requires a matching SEAD context window.

Sensitivity rank is score-ordered for isolating weight effects; the published engine rank separately applies direct-evidence tie-breakers.

The maximum absolute movement across the tested profiles is 0 ranks.

Archaeology-Emphasis Top 20

Rank	Lake	Score	Archaeology signal at 20 km	SEAD time windows	Numeric SEAD records	Shift from baseline