MODERN ANALOGS IN QUATERNARY PALEOECOLOGY: Here Today, Gone Yesterday, Gone Tomorrow?

Modern analog analysis, the comparison of Quaternary fossil pollen assemblages with modern assemblages, has long been a mainstay of paleoecological and paleoclimatic inference. The logic of analogical inference involves a comparative element (comparison of modern and fossil assemblages to select matches and assess goodness of fit) and a causal element (assumption that the relationships between modern vegetation and derivative pollen assemblages are matched by those between ancient vegetation and fossil pollen assemblages). An array of numerical and statistical tools have been developed to ensure objective, consistent, and quantitative assessments of similarity between pollen assemblages. Divergent or convergent relationships between vegetation and pollen assemblages can arise from a variety of sources, composing a potential source of error in analog analysis, but such errors can be anticipated and minimized. Pollen assemblages lacking modern analogs are well documented for the late-glacial period (17,00010,000 years BP) in eastern North America and other regions. Simulated climates for this period also lacked modern analogs owing to increased seasonality of insolation, lowered CO2, and persistent ice sheets. Most pollen assemblages from the last glacial maximum (23,00020,000 years BP) in eastern North America have modern analogs, but macrofossil and other evidence suggest that the vegetation may have lacked modern analogs, owing to unique climate realizations and perhaps direct effects of lowered CO2. Better understanding of the nature of past no-analog vegetation, and the underlying causes, will address important issues in ecology and evolutionary biology and help anticipate biotic responses to the no-analog greenhouse world of the near-future.