Paleofidelity: An R Package for Measuring and Visualizing Live-Dead Fidelity

Abstract

Live-dead fidelity analysis, one of the key approaches of conservation paleobiology, aims to measure the congruence between living communities and sympatric death assemblages. Typically, data involve compositional matrices with counts of specimens grouped by variables (typically taxa) and observations (typically sampling units). The most common targets of live-dead comparisons are diversity fidelity (especially alpha diversity and evenness) and compositional fidelity (faunal agreement). A beta version of a new package dedicated to analyzing and visualizing live-dead fidelity is available on GitHub (R package: “PaleoFidelity”). The package allows for measuring diversity fidelity (dead-live offset in sample-standardized species richness), evenness fidelity (dead-live offset in Hulbert’s Pie evenness), and compositional fidelity estimated by correlation (Spearman, Kendall, and Pearson) and similarity (Bray, Chao, etc.) measures. The package includes a resampling model for assessing expected values of correlation and similarity measures under the null model of perfect live-dead congruence. In addition, tests and confidence intervals based on resampling protocols are provided to allow for statistical assessment of fidelity patterns. Finally, PaleoFidelity includes plot functions for visualizing live-dead congruence in diversity or faunal composition. The current version of the package can be installed in R or R Studio using the following statement: devtools::install_github(‘mjkowalewski/PaleoFidelity’, build_vignettes = TRUE).