FRANK L. FORCINO, EMILY S. STAFFORD, JARED J. WARNER, AMELINDA E. WEBB, LINDSEY R. LEIGHTON, CHRIS L. SCHNEIDER, TOVA S. MICHLIN, LAUREN M. PALAZZOLO, JARED R. MORROW, STEPHEN A. SCHELLENBERG
PALAIOS 25 (3), 144-157, (1 March 2010) https://doi.org/10.2110/palo.2009.p09-011r
Paleocommunity research efforts have explored a multitude of faunal assemblages using a wide range of sampling and analytical methods to infer a paleoecological signal. Here, we derive six secondary datasets from a single stratigraphic series of faunal assemblages in the Finis Shale (Pennsylvanian) of Jacksboro, Texas, USA, using a variety of data categorization decisions (i.e., abundance versus calcified biomass, all taxa versus selected indicator taxa, and generic versus higher clade resolution). Biomass- and abundance-derived datasets were not significantly different in terms of evenness, Shannon's information index, or Simpson's diversity index. Using Bray-Curtis and nonmetric multidimensional scaling ordinations, with Sorenson and relative Sorenson distance measures, ordination axis scores of the six derived datasets were all significantly correlated with one another, suggesting little difference in their respective paleoecological signals. Three potential explanations for this consistent paleoecological signal, regardless of which data categorizations are employed, include: (1) the dominance of a few brachiopod taxa overwhelmingly influenced the community structure, (2) relatively constrained environmental conditions limited community variation, and (3) low variation in specimen size minimized potential differences among abundance and calcified biomass categorizations. We suggest that other datasets with greater diversities, greater evenness, or from a wider range of paleoenvironments might not show this consistency. Thus, to the degree possible and appropriate, paleoecological investigators should test the effects of these data categorization decisions on a paleoecological signal, regardless of the analytical method employed.