Fig 1.
Pollen-based reconstructed tree cover at LGM.
Estimated tree cover at LGM (APFT%), calculated as percent sum of all PFT scores belonging to arboreal PFTs. The tree cover surface is based on the 3D interpolation of PFT scores from individual fossil pollen sites (circles). Method following [6]. For data sources see S3 Table.
Fig 2.
Archaeological evidence for human population in LGM Europe.
(A) Distribution of archaeological sites dating to LGM [17, 18], (B) estimated baseline population density using kernel interpolation following [19].
Fig 3.
Forager population density and human induced fire at LGM.
(A) Forager population density, and (B) the change in mean annual burned area from the LPJ-LMfire ensemble mean scenario.
Fig 4.
Influence of anthropogenic burning on tree cover in LGM Europe.
(A) Tree cover simulated without and (B) with human burning. Results are from the LPJ-LMfire ensemble mean scenario.
Fig 5.
Change in burned area fraction between LGM and the Preindustrial Holocene (PIH).
Difference in mean annual burned area fraction between LGM and the late PIH (AD 1770) in (A) the LPJ ensemble mean simulation without human burning, and in (B) the LPJ ensemble mean scenario including human burning. Panel (C) shows relative charcoal accumulation at the LGM and late PIH (AD 1650±200 yr) at the five Global Charcoal Database sites (D) containing lake and marine sediment microcharcoal dating to LGM [23]. The Z-scores of transformed charcoal influx are calculated relative to the preindustrial Holocene base period (0.2–12 ka) following [21]. At all charcoal sites the charcoal influx is smaller in samples dated to LGM compared to the PIH. The gray outline is the LGM coastline.