PALEODATA-MODEL COMPARISONS OF ARCTIC VEGETATION RESPONSE TO CLIMATE AT 6 AND 21 KA

Arctic vegetation responds strongly to climatic factors such as growing season warmth and moisture availability. We investigated the response of vegetation to past climate using a data-model comparison approach. We developed a circumpolar classification, based on the abundances of plant functional types, of five tundra and seven forest vegetation types. This drove a ‘biomization’ procedure, which converts pollen assemblages to vegetation cover, for arctic pollen data for 0, 6, and 21 ka. An equilibrium biogeochemistry - biogeography model (BIOME 4), which uses the same vegetation classification driven by climatology, was coupled with GCM paleo-simulations for 6 and 21 ka.

Biomized and simulated vegetation for 0 ka match mapped data well in most sectors. Data for 21 ka show grass-forb tundra and prostrate/dwarf-shrub tundra dominating unglaciated Alaska and east/central Siberia; forest and shrub tundra cover is greatly reduced. At 21 ka, the models tend to generate too much precipitation, with only one simulating conditions dry enough to capture the full observed reduction of shrub tundra. At 6 ka, both data and model simulations show that the northward extension of treeline related to insolation forcing of warm summers is muted and asymmetric, with advances in central Siberia but no advance in Beringia. The asymmetry is attributed to Siberia’s greater continentality and to sea-ice dynamics. In northeast Canada, a southward shift in treeline recorded by the data is counter to a simulated northward shift; the discrepancy is likely due to the absence in the models of residual Laurentide ice and the simulation of an associated regional cooling.