Examining climate and composition differences in eastern North American temperate forests using pollen ratios and pollen assemblage cluster analysis

Ecologically and economically important temperate forests cover much of North America, and these forests contain compositionally distinct forest subtypes. The relationships between vegetation composition, climate, and the pollen taxa that represent eastern North American temperate forests have been studied for decades. However, few have examined the pollen taxa that represent subtypes of these forests at the ecoregion level. Pollen ratios of key taxa and self-organized mapping (SOM) analysis both show promise for identifying compositional differences in vegetation, and ultimately, climate. Using contemporary pollen samples from surficial lake sediments in the western Great Lakes region, we evaluated both methods. Quercus:Fagus, Quercus:Betula, Carya:Pinus, and Carya:Betula pollen ratios were tested for their ability to reveal relative levels of mean annual precipitation, mean annual temperature, and mean growing season precipitation, as well as to distinguish between ecoregions. All four ratios were significantly positively correlated with almost all tested climate metrics and were able to differentiate between ecoregions in many instances. We suggest that the Quercus:Betula ratio shows the most promise for application in palaeoecological studies because the two taxa have opposite spatial distributions and incorporate species that occupy more specific climate spaces relative to the other ratios. Additionally, the SOM analysis revealed three sample groups that cluster geographically and appear to represent distinct forest subtypes. Based on these results, SOM has potential for application in future palaeoecological reconstructions.