Deriving ecological and evolutionary descriptions of, and implications from, faunal assemblage patterns is commonly addressed by observation and a variety of exploratory techniques (scaling and clustering), along with qualitative evaluations of species occurrences and relative abundances. We argue that interpretations of faunal patterns, especially those documented by the fossil record, should be based upon the composition and structure of entire communities to provide strong conclusions and replicable results.

As an example, we use benthic foraminiferal data at high resolution (1–2 cm, corresponding to 300–1400 yr) over a section corresponding to about 20 kyr across the beginning of the Paleocene–Eocene thermal maximum (PETM). The PETM was an episode of rapid global warming about 55.5 Ma, associated with ocean acidification and lowered open oceanic productivity and deoxygenation and marked by severe turnover in benthic foraminiferal assemblages. Here we provide a stand-alone approach applicable to any dynamic faunal system, perturbation detection analysis (PDA), to recognize and identify community disruption evidenced as either positive growth or negative decline, and we use this methodical approach to obtain new information on foraminiferal communities before, during, and after the initiation of the PETM.

We conclude that the late Paleocene benthic foraminiferal community (FCOM1) was in a growth stage of positive increasing diversity, suggestive of favorable environmental conditions. This stage continued through the initial changes at the onset of the PETM, when disruption through environmental stress led to this community's termination. A second community (FCOM2) formed with declining diversity and high variability, showing a lack of adaptation to changing conditions. Knowledge of total assemblage status under both adverse and advantageous conditions is necessary, but not recognized by methods that rely upon analysis of single samples only: individual samples cannot be used to recognize disruptive changes in a community's structure, but these are easily identified using PDA.