Replication Data for: Palaeoecological signals remain robust at broad spatiotemporal scales despite artificial reductions in sampling power and geographical scope

This repository contains the data and code to replicate the results for 3.1, 3.2, or 3.3 in the paper titled "Palaeoecological signals remain robust at broad spatiotemporal scales despite artificial reductions in sampling power and geographical scope." It also contains the citations for the data used in the analysis. Project summary: Palaeoecological analyses rely on datasets that are comprised of collections or observations of fossil specimens. When palaeoecologists consider the scope of a novel project or research question, logistical constraints surrounding specimen acquisition are often a primary limiting factor. Despite this, some exceptional datasets exist that are both broad in spatiotemporal scope and densely sampled. One such dataset, consisting of 116,896 fossil brachiopods from central Nevada spanning from the beginning of the Early Devonian to the end of the Middle Devonian, is utilized by the present study as a model dataset to test the effects of logistically imposed restrictions on sample collection. Specifically, the present study tests whether palaeoecological signals (as represented in an NMDS ordination plot), and the processes that drive them, may be identified despite artificial reductions in spatiotemporal resolution and scope. These artificial reductions simulated two types of logistical sampling constraints: (1) limitations on the absolute number of samples collected, and (2) inability to travel to, and collect from, geographically separated regions. The primary palaeoecological signal identified in NMDS ordination plots of the entire dataset (255 samples) was identifiable more than >98% of the time with as few as 26 randomly selected samples, and NMDS ordinations of geographically restricted subsamples representing temporal intervals of >5 Ma were found to contain signals representative of the entire study area. Thus, it was possible to capture the primary signals corresponding to biodiversity shifts on broad scales with limited sampling power, and without extensively sampling outside of a single geographical region.