Abstract Detail

Macroevolution

Simpson, Andrew [1], Korasidis, Vera [2].

On the use of databases in paleobotanical research: triumphs, pitfalls, and good and bad examples.

Online databases of botanical records of both living and fossil plants provide opportunities for researchers to conduct broad-scale analyses without the time, funding, permission, and manpower constraints required for fieldwork or new analysis from existing collections. However, the results of studies reliant upon databases are only as trustworthy as the data incorporated, and scientists performing multidisciplinary studies unfamiliar with the pitfalls of a particular dataset may draw incorrect conclusions because of said unfamiliarity. Paleobotanical data are often different from neobotanical data in several notable ways: first, organs (i.e., leaves, infructescences, fruit, pollen) are often preserved separately, and consequently given separate taxonomic names. Incorporating each organ individually in measures of diversity could lead to overestimations of diversity in epochs/successions where the taxa have been well described. Second, paleobotanists are more confident assigning younger (i.e., Cenozoic) fossils to living taxa than older (i.e., Palaeozoic or Mesozoic) fossils; consequently, older fossils often have different names from morphologically identical younger fossils. Modern paleobotanists are also now far less willing to assign fossils to living genera based purely on gross visual similarity than past paleobotanists. Third, again because of uncertainty surrounding morphology, similar fossils collected from different geographic regions representing the same plant, can be given different names, possibly artificially inflating beta and gamma diversity.
We here review a selection of studies from the published literature that use the Paleobiology Database (PBDB) as a source of paleobotanical data in synthetic analyses. Some of the authors of these papers applied appropriate caution concerning the aforementioned pitfalls and thus produce trustworthy (from a paleobotanical prospective) results. Others do not consider the biases imposed by paleobotanical data and take the PBDB records at face value, and thus produce misleading results, among them extinction and radiation events that are not real, temporal ranges of fossil taxa in excess of what is probable, and other issues. We summarize the problems that arise from these misunderstandings of paleobotanical data, and make suggestions for future scientists, both for neobotanists interested in using paleobotanical databases, and for managers of databases containing paleobotanical data, that might reduce confusions in the future.

1 - Department Of Paleobiology, National Museum Of Natural History, Smithsonian Institution, Washington, DC, 20560, United States
2 - Smithsonian Institution, Paleobiology, Washington, DC, USA

Keywords:
Databases
Paleobotany
Meta-analysis
Review
Bias.

Presentation Type: Oral Paper
Session: MACROI, Macroevolution I
Location: Virtual/Virtual
Date: Tuesday, July 20th, 2021
Time: 1:30 PM(EDT)
Number: MACROI005
Abstract ID:159
Candidate for Awards:None