Multivariate methods for testing hypotheses of temporal community dynamics

Abstract

For ecological research to make important contributions towards understanding and managing temporally-variable global change processes, such as responses to land-use and climatic change, we must have effective and comparable ways to quantify and analyse compositional change over time in biological communities. These changes are the sum of local colonisation and extinction events, or changes in the biomass and relative abundance of taxa within and among samples. We conducted a quantitative review of currently available methods for the analysis of multivariate datasets collected at temporal intervals. This review identified the need for the application of quantitative, hypothesis-based approaches to understand temporal change in community composition, particularly for small datasets with less than 15 temporal replicates. To address this gap, we: (1) conceptually present how temporal patterns in community dynamics can be framed as specific, testable hypotheses; (2) provide three fully-worked case-studies, complete with R code, demonstrating multivariate analysis methods for temporal hypothesis testing and pattern visualisation; and (3) present a road map for testing specific, quantitative hypotheses relating to the underlying mechanisms of temporal community dynamics.