Temperature variations in the southern Great Lakes during the last deglaciation: Comparison between pollen and GDGT proxies

Abstract

Our understanding of deglacial climate history in the southern Great Lakes region of the United States is primarily based upon fossil pollen data, with few independent and multi-proxy climate reconstructions. Here we introduce a new, well-dated fossil pollen record from Stotzel-Leis, OH, and a new deglacial temperature record based on branched glycerol dialkyl glycerol tetraethers (brGDGTs) at Silver Lake, OH. We compare these new data to previously published records and to a regional stack of pollen-based temperature reconstructions from Stotzel-Leis, Silver Lake, and three other well-dated sites. The new and previously published pollen records at Stotzel-Leis are similar, but our new age model brings vegetation events into closer alignment with known climatic events such as the Younger Dryas (YD). brGDGT-inferred temperatures correlate strongly with pollen-based regional temperature reconstructions, with the strongest correlation obtained for a global soil-based brGDGT calibration (r2 = 0.88), lending confidence to the deglacial reconstructions and the use of brGDGT and regional pollen stacks as paleotemperature proxies in eastern North America. However, individual pollen records show large differences in timing, rates, and amplitudes of inferred temperature change, indicating caution with paleoclimatic inferences based on single-site pollen records. From 16.0 to 10.0ka, both proxies indicate that regional temperatures rose by ∼10 °C, roughly double the ∼5 °C estimates for the Northern Hemisphere reported in prior syntheses. Change-point analysis of the pollen stack shows accelerated warming at 14.0 ± 1.2ka, cooling at 12.6 ± 0.4ka, and warming from 11.6 ± 0.5ka into the Holocene. The timing of Bølling-Allerød (B-A) warming and YD onset in our records lag by ∼300–500 years those reported in syntheses of temperature records from the northern mid-latitudes. This discrepancy is too large to be attributed to uncertainties in radiocarbon dating, and correlation between pollen and brGDGT temperature reconstructions rules out vegetation lags as a cause. However, the YD termination appears synchronous among the brGDGT record, regional pollen stack, and Northern Hemisphere stack. The cause of the larger and lagged temperature changes in the southern Great Lakes relative to Northern Hemisphere averages remains unclear, but may be due to the effects of continentality and ice sheet extent on regional climate evolution.