Assessment of assemblage‐wide temporal niche segregation using null models

Abstract

Summary 1. Although time can be subdivided to promote species coexistence, quantitative examination of assemblage‐wide temporal niche overlap has seldom been explored with appropriate null models. Because of the sequential and continuous nature of time, it requires a different kind of randomization model than those used to assess subdivision of discrete and non‐sequential resources (e.g. food types and microhabitats). 2. For two common niche overlap indices (Pianka and Czekanowski), we compared the responses of two common randomization models and a newly developed model (ROSARIO) to different levels of temporal autocorrelation, specialization and coincidence of activity. 3. Although qualitatively similar results characterized overlap indices, results differed depending on randomization model. Temporal resolution of the data and amount of temporal specialization in an assemblage can have large effects on model outcomes. ROSARIO is as powerful as the models used for analyses of overlap of nominal and unordered resources, but it is more appropriate for ranked and interval data, as it maintains the empirical temporal autocorrelation within species. 4. ROSARIO can be a useful tool for exploration of assemblage‐wide patterns of overlap in the use of resources that occur as cyclical phenomena, such as diel phases, yearly seasons, lunar tides and climate oscillations.