Distribution of diatoms among intermittent ponds on the Atlantic Coastal Plain: development of a model to predict drought periodicity from surface-sediment assemblages

• Published in: Journal of paleolimnology 1998-07, Vol.20 (1), p.71-90
• Main Authors: Gaiser, EE, Philippi, TE, Taylor, B E
• Format: Article
• Language: English
• Subjects: Bacillariophyceae; Freshwater
• ISSN: 0921-2728
• DOI: 10.1023/A:1007969500673

Inferences of past climate from the fossil record in lakes rely on the accurate quantification of a relationship of fossilizing organisms to their environment. Whereas the relationship of diatoms to water chemistry parameters has been modeled in many systems, few studies adequately address the relationship of diatoms to physical properties, such as water depth or hydrology, that may be more directly tied to climate. We examined the composition of modern diatoms in surface sediments of 75 isolated ponds (mostly Carolina bays) of the Atlantic Coastal Plain to: (1) assess the influence of physical and chemical variables on the distribution of diatoms among ponds of the region, and (2) develop a model that predicts hydroperiod (a measure of pond permanence) from diatom assemblages. We constructed two hydroperiod calibration models: the first infers hydroperiod from the weighted-average optima and tolerances of taxa along the hydroperiod gradient, the second bases inferences on the hydroperiod estimates of compositionally similar samples. Both approaches incorporate a-priori and post-hoc tests of assumptions often inherent in the construction of transfer functions. Diatom assemblage composition had strong, approximately linear relationships to hydroperiod, water depth, and calcium concentration in non-metric multidimensional ordination space; effects of other variables, including pH, were non-linear or ambiguous. Overall, the assemblages reflected the dilute, acidic chemical characteristics of bays. The assemblages contained differing abundances of euterrestrial, benthic and planktonic taxa, depending on a pond's susceptibility to drying. A weighted-averaging regression model based on taxon-specific hydroperiod optima generated adequate, unbiased hydroperiod inferences from diatom species composition (r super(2) = 0.81). This model may be used to infer past drought episodes from fossil diatom assemblages at appropriate sites on the Atlantic Coastal Plain.