SAMPLING COMPLETENESS IN THE NORTH AMERICAN PHANEROZOIC MARINE FOSSIL RECORD: PALEOZOIC CARBONATE OVERSAMPLING AND POTENTIAL SOURCES OF BIAS

BALSEIRO, D.; POWELL, M.G.

Congreso; Congreso de la Asociación Paleontológica Argentina; 2021

Paleontologists have always emphasized efforts to understand how sampling the fossil record can bias our understanding of the history of life. Most studies, however, limited the analysis to comparisons between sampled diversity and some proxy of rock availability. But diversity is not a good proxy for sampling effort because it can vary for biological reasons as well as sampling. To overcome this problem sampling units need to be directly linked to the rock record. We here integrate North American marine invertebrate fossil occurrences from the Paleobiology Database with lithologic information from Macrostrat to reveal sampling patterns in marine carbonates and siliciclastic rocks. We aim to understand temporal trends in sampling effort within and between lithologies focusing on the proportion of total available volume that has been sampled (i.e., sampling completeness). Results indicate that total sampling completeness is stable during the Paleozoic and very variable during the post-Paleozoic, but showing no increase through time. Carbonate completeness relative to siliciclastic completeness (i.e., excess in carbonate completeness) shows two intervals with contrasting sampling patterns. Paleozoic carbonates are invariably sampled more extensively than siliciclastics, with intervals where carbonate completeness doubles siliciclastic completeness. Post-Paleozoic rock record shows a much variable sampling pattern with siliciclastic completeness being higher relative to carbonate completeness in many intervals. Among possible explanations, some authors have proposed that the high frequency of Paleozoic sedimentary units barren of fossils could be a real pattern caused by environmental harshness. However, an analysis of potential fossiliferous volumes indicates that frequent widespread anoxia in the Paleozoic is an implausible explanation for siliciclastic rocks being less fossiliferous than carbonate rocks. A taphonomic origin for the described sampling pattern was tested comparing carbonate completeness of calcitic and aragonitic fossils. Similar sampling patterns suggest that early dissolution of aragonitic shells in siliciclastic environments is not the cause of the registered carbonate oversampling. Further analysis indicated that problems in data entry in the Paleobiology Database cannot account for a major potential source of bias. A last possibility could be that paleontologists publish fossils from carbonate rocks more frequently than from siliciclastic rocks. The higher diversity and evenness of occurrence distributions in carbonate rocks might account for an ease in the description of unknown species and therefore higher rate of published fossils coming from this lithology. Finally, a strong bias in favor of carbonate rocks can distort our perception of diversity even under commonly used standardization methods.