Turnover in the Devonian Foreland Basin of New York: How Geographic Coverage Affects Our Understanding of Origination and Extinction Rates

Variations in sampling intensity and coverage have the potential to greatly modify our understanding of biodiversity, origination, and extinction through time. As the sampled area increases, a wider variety of habitats and a greater number of individuals are likely to be sampled. Poorly understood, however, is how shifting geographic ranges may affect calculated biodiversity metrics independent of sampling. If organisms or groups of organisms migrate in and out of the geographic sampling window in response to changes in their environment, origination and extinction rates could be greatly distorted. To help assess the impact of changes in geographic range of taxa, we are constructing a time-space database of Devonian fossil occurrences for New York State. Known occurrences of taxa are combined with geographic and stratigraphic data, allowing us to assess how diversity varies in both spatial and temporal dimensions. The dataset currently consists of over 10,000 generic occurrences from over 450 collections. Because the outcrop belt runs roughly east-west, a north-south (longitudinal) slice through the state captures localities that span the entire Devonian. We repeatedly subsample the data by selecting random longitudinal windows of variable width, and then calculate rates of origination and extinction for each window. Origination and extinction rates are highly variable at small geographic scales but variance rapidly decreases with increasing longitudinal bin size, stabilizing when roughly 25% (1 degree longitude) of the data are sampled. Interestingly, extinction rates appear to be better buffered against changes in geographic coverage than origination rates. Although broad trends are similar in different formations, important differences exist as to the precise bin size at which origination and extinction rates stabilize. These discrepancies might be explained by differences in the areal extent of outcrop, number of samples, or beta diversity between time bins.