Null model approaches to evaluating the relative role of different assembly processes in shaping ecological communities

Abstract

Various local processes simultaneously shape ecological assemblages. β-diversity is a useful metric for inferring the underlying mechanisms of community assembly. However, β-diversity is not independent of γ-diversity, which may mask the local mechanisms that govern community processes across regions. Recent approaches that rely on an abundance-based null model could solve this sampling issue. However, if abundance varies widely across a region, the relative roles of deterministic and stochastic processes may be substantially misestimated. Furthermore, there is additional uncertainty as to whether null models used to correct γ-dependence in β-diversity should be independent of the observed patterns of species abundance distributions or whether the models should reflect these patterns. Here, we aim to test what null models with various constraints imply about the underlying processes shaping β-diversity. First, we found that an abundance-driven sampling effect could substantially influence the calculation of γ-corrected β-diversity. Second, we found that the null models that preserve the species abundance patterns could better reflect empirical patterns of spatial organization of individuals. The different implications generated from different applications of the null model approach therefore suggest that there are still frontiers regarding how local processes that shape species assemblages should be quantified. Carefully exploring each facet within different assembly processes is important.

Resource availability determines the importance of niche‐based versus stochastic community assembly in grasslands

“…The probability of a species to be drawn was based on the proportion of plots occupied by this species across all treatments (Chase et al ). Constraining sampling probability is often done in null‐models to prevent rare species of becoming unrealistically frequent and vice versa (Mori et al ). The number of random draws in which the expected number of shared species is larger than the observed one is then divided by the total number of random draws (n = 9999).…”

Integrating species traits into species pools

“…assembly including null-model analyses of spatial or temporal variation in community composition, species-environment or trait-environment relationships, variance partitioning of spatial and environmental factors, and/or analyses of trait over-/under-dispersion in local communities (Legendre et al 2005, Brown et al 2014, Swenson 2014, Mori et al 2015, Ovaskainen et al 2017.…”

How Should Beta-Diversity Inform Biodiversity Conservation?

“…However, new techniques using the distance-decay of pairwise similarity provide a potentially powerful tool for extrapolating small studies to larger landscapes [44](Box 3). 100 Even when the scaling factor is measured appropriately, the conservation significance of a change in beta-diversity is not straightforward. Maximizing betadiversity is not necessarily desirable for gamma-diversity conservation, because damaging anthropogenic impacts can cause the similarity of local communities to increase, decrease, or remain unchanged, depending on the relative balance of 105 homogenization and heterogenization processes at the site level ( Figure 1).…”