Background/Question/Methods A large decline in beech populations (Fagus grandifolia) has been documented from pollen records in southeastern Michigan and Southern Ontario between 1000 and 600 BP. However, no major vegetation changes are observed in pollen records from the lake-effect snow belt of Upper Michigan during the same period. Declines in beech pollen were generally associated with increases in oak (Quercus) and pine (Pinus). The causes of these forest changes and their spatial patterning are not completely understood, although they have been variously attributed to anthropogenic disturbance, climatic cooling, or drought. We have conducted tandem investigations of pollen, charcoal, and hydroclimate proxies (testate amoebae, humification) from the archives contained in four Sphagnum peatlands of the Great Lakes region to test the hypothesis that drought and fire caused the forest changes at this time. Our site network is also designed to test the hypothesis that increased lake-effect snow may have buffered some regions from the effects of drought. Two of our studied peatlands were located within the lake effect snow belt (northern MI, northwestern IN) and two were located outside the lake effect snow belt (northeastern MI, southeastern MI).

Results/Conclusions Our results reveal that a series of large droughts and fires were associated with the beech decline at both sites in eastern Michigan. At these sites, high-resolution analysis of beech pollen and hydroclimate proxies reveals complex ecological dynamics at multidecadal timescales between 1000-600 BP, with large fluctuations in beech pollen, available moisture, and charcoal concentrations. However, at sites in the lake-effect snow belt of Upper Michigan and northwestern Indiana, no droughts or decline in beech populations were recorded. Our study demonstrates the potential for integrating paleoecological data with vegetation-independent paleoclimate proxies to assess ecological responses to climate variability at multidecadal timescales.