Reconstructing historical ranges with fossil data at continental scales

Abstract

Sedimentary records indicating the changing distribution of temperate tree species since the last ice age are widely used to understand the rates and patterns of population response to environmental change and the genetic consequences of such dynamics. It is well known that fossil pollen and plant macrofossil data provide limited information on distribution and abundance of small or diffuse tree populations. Nevertheless, the estimates of potential migration rates that are currently in use rely on assumptions about how fossil pollen and macrofossils record such populations. To better understand how such assumptions may influence estimates of spread rate, we related modern tree distributions to surface pollen abundance, using assumptions that are routinely applied to paleoecological data. We were unable to reconstruct the modern ranges of two species that are well represented in sediments using pollen and macrofossils from sediment surface samples. The ranges of American beech ( Fagus grandifolia) and yellow birch ( Betula alleghaniensis) include large areas where these trees are not abundant, and these were impossible to identify using pollen and macrofossil data, respectively. Our analysis supports the view that populations of these and other species existed at low densities over extensive areas in the past, and such areas would not be accurately mapped using fossil data. If so, estimates of past migration rates could be inaccurate.