ON THE INTERPRETATION OF FOSSIL POACEAE POLLEN IN THE LOWLAND HUMID NEOTROPICS

Abstract

Poaceae pollen is abundant in fossil records and is often used as a paleoclimatic indicator. A common interpretation is to link increases in Poaceae pollen abundance to increased regional aridity. However, the representation of Poaceae pollen is influenced by a number of factors, such as the proportion of other plants in the flora that are anemophilous, the size of local marshes, and the influence of humans on the landscape. Abundant anemophilous trees are likely to mask the contribution of Poaceae pollen, whereas floras that are primarily entomophilous are likely to produce a pollen spectrum containing an over-representation of Poaceae. As most fossil pollen data are drawn from flooded settings, it is critically important that palynologists recognize the Poaceae pollen contribution derived from floating grasses and marshes that surround their coring site. Interpretations that ignore changes in effective lake size and assume that Poaceae percentage is a simple indicator of regional vegetation change are likely to overstate 'dry' episodes and transitions from wet forest to scrub environments. Human occupation of a site is sometimes manifested in increased Poaceae pollen abundance in lake sediments. The duration, intensity and land-use associated with occupation are all variables that influence Poaceae pollen representation. Very high (50-90%) abundances of Poaceae pollen provide a strong indicator of savanna habitats, but trying to determine transitional vegetation types between savanna and wet forest is best determined using other taxa. Indeed, reliance on Poaceae abundance as an indicator of paleoprecipitation is potentially very misleading when it is in the fossil record at moderate abundances.