The influence of dataset decisions on paleocommunity results: Comparison of different categorizations of two datasets from Devonian carbonate units

Community paleoecology researchers have employed an array of methods for collecting, counting, and identifying fossils. Few studies have quantified the differences among results when employing various methods on a single stratigraphic series of paleocommunities. Here, we examined fossils collected from the Rapid Member of the Little Cedar Formation of Coralville, Iowa, USA and the Waterways Formation of northern Alberta, Canada to compare the paleocommunity results between the use of abundance versus point counts as a counting method and between using all taxa in a paleocommunity versus only the brachiopod taxa. Comparisons between abundance and point count categorizations resulted in goodness-of-fit statistics ranging from 0.61 to 0.92. In addition, we visually assessed each of the ordination comparisons and interpreted each categorization comparison as leading to different paleocommunity interpretations. Colonial and easily disarticulated taxa (i.e., crinoids and bryozoans) had a strong influence on the differences between abundance and point count categorizations; when the colonial and easily disarticulated taxa were culled and the analyses rerun, goodness-of-fit statistics increased. When possible, we recommend the use of both abundance and point counts when conducting paleocommunity analyses. However, because point counts allow the equal weighting of solitary, non-solitary, and easily disarticulated taxa, point counts may be a more accurate means of measuring taxonomic composition of paleocommunities. Furthermore, because of the disparity between the paleocommunity results of brachiopods-only versus all-taxa, we recommend the use of all taxa for environmental, ecological, or evolutionary research based on paleocommunities.