Recent advances in metacommunities and meta-ecosystem theories

Communities result from the balance between sorting and movement

One of the main predictions of metacommunity theory is the regional distribution of species diversity based on both local competition and regional movement of species. In competitive metacommunity models, sorting, or environmental filtering, is the local process that operates through competitive exclusion, and sorting can lead to regional patterns of diversity through spatial heterogeneity in the environment. Movement among local communities is the regional process that redistributes, and eventually homogenizes, local communities, working against local sorting. Sorting is purely local in the absence of movement, whereas strong movement leads to regional sorting that favors the best competitor in the average habitat¹¹. In spatially implicit models assuming a homogeneous (all-to-all) connectivity network, intermediate movement reveals the reciprocal effects of local and regional processes: both regional and local diversity and relative species abundance then result from the balance between local sorting and immigration from the regional metacommunity. This balance can, for example, predict the contributions of both local species richness (alpha diversity) and regional species turnover (beta diversity) to community stability, through local compensatory effects, and to the maintenance of regional asynchrony in species composition¹².

The balance between local and regional processes in heterogeneous metacommunities was used to propose four mechanisms controlling biodiversity in competitive metacommunities³ based on the relative importance of species movement and habitat heterogeneity as driving processes (species sorting, patch dynamics, mass effect, and neutral). The proposed mechanisms implicitly assumed demographic stochasticity, which stimulated the integration of a metacommunity framework to the neutral-versus-niche debate. Demographic stochasticity was treated explicitly as drift, affecting community dynamics along with competition, dispersal, and speciation^13,14. Recent studies have provided a coherent framework reconciling competition and metacommunity theories¹⁵, whereas others have refined statistical methods for the inference of neutral- and niche-based mechanisms of species diversity¹⁶. Another important integration has recently happened with theories of biodiversity-ecosystem functions, tying into a long-standing empirical debate about the role of species richness for productivity and stability¹⁷. Within that debate, the competitive metacommunity framework can reconcile apparent contradictions in observed biodiversity-ecosystem function relationships. Dispersal limitation can, for example, either prevent or saturate local sorting processes and affect the strength of species complementarity underlying the stability and productivity of local communities⁸. Metacommunity dynamics can also resolve the combined and sometimes opposite effects of species diversity on multiple ecosystem functions¹⁸.

Spatially explicit metacommunities have blurred the distinction between local and regional scales because spatiotemporal heterogeneity in the distribution of species can emerge at multiple scales rather than being a priori-defined. Also, when the local scale approaches the individual level, local and regional (meta)communities become arbitrarily defined along a continuum, based on assumptions about the relevant scales of processes at both ends of the continuum: competitive exclusion in local communities versus speciation and immigration over the metacommunity¹⁹. This is illustrated by the application of the metacommunity framework to neutral theory. In early neutral models, the metacommunity was independent from the local community, ignoring feedbacks from local to regional levels^20,21. Later studies integrated both scales through spatially explicit models where metacommunity processes emerge from the collective behavior of connected local communities²². More recently, this scale continuum has been used to improve the testability of neutral predictions using patterns of spatial autocorrelation²³.

The competitive metacommunity framework emphasizes spatial heterogeneity among communities and as the regional mechanism of coexistence. It has also integrated the role of temporal environmental variability as a mechanism of coexistence^24,25. However, the emphasis on sorting through competitive interactions has limited the understanding of other traits and interactions, and of their complexity, as drivers of species diversity and function. The study of trophic interactions allowed researchers to extend this understanding beyond the effects of competitive exclusion to networks of species interactions that are more compatible with the complexity of natural systems.

Trophic metacommunities: networks of networks

The extension of competitive metacommunity theories to trophic metacommunities leads to a ‘spatial network of interaction networks’²⁶. One major implication of trophic metacommunities is the generalization of species traits involved in the local and regional control of community structure and dynamics. This generalization has recently led to a deeper contribution of metacommunities to the community complexity-stability relationship, involving the integration of evolutionary and non-equilibrium dynamics to the framework.

Trophic interactions have expanded the exploration of trait differentiation at both local and regional scales: body size²⁷, trait-mediated and density-dependent movement²⁸, or covariance between trophic position and movement^29–31. It has recently led researchers to revisit the hump-shaped relationship between dispersal and diversity in competitive metacommunities. When multiple trophic levels with heterogeneous dispersal rates are considered, this expected balance between immigration and sorting can instead give rise to monotonic increase in diversity with dispersal³⁰ (Figure 1A). The dependence of movement across trophic levels also provides a trophic context mediating local (risk-based movement) and regional (perceived habitat suitability) drivers of species diversity³². Similarly, density-dependent movement in heterogeneous tri-trophic metacommunities was shown to best capture patterns of beta diversity and, more specifically, the decrease in similarity with distance observed in natural systems²⁸.

Figure 1. Role of spatial interaction strength across trophic levels for dispersal-diversity relationships and for regional stability.

(A) Diversity–dispersal relationships. (a) Consumer dispersal α varies while keeping resource dispersal β small. (b) Consumer dispersal α and resource dispersal β vary simultaneously. (c) Resource dispersal β varies while keeping consumer dispersal α small. Thick line: local diversity; thin line: regional diversity. Reprinted with permission from John Wiley & Sons, Inc.³⁰. (B) Changes in qualitative dynamics in a two-patch tri-trophic meta-ecosystem from varying resource intra-specific interaction strength (β) and dispersal rates, holding all other parameters constant. At intermediate levels of β, making dispersal more non-hierarchical by decreasing middle-trophic-level dispersal rates (d_H) leads to increased stability because of the emergence of stable asymmetric equilibria. Reprinted with permission from the University of Chicago Press Books³¹.

Food-web metacommunities have explored the assembly of large networks of networks as steady-state structures^33,34 that drive their own stability³⁵ and productivity. This ‘network of networks’ approach provides a useful framework to integrate dispersal limitation, habitat heterogeneity, and food-web topology^26,36,37 to the complexity-stability debate initiated by May³⁸. For example, May’s predicted limit to the complexity of stable food webs could be relaxed by increasing habitat heterogeneity across metacommunities characterized by intermediate dispersal^26,37. Moreover, Pillai et al.³⁶ showed how the interaction between dispersal and species interactions would predict the stabilizing effect of omnivorous and generalist species, which adds to predictions from non-spatial food-web theories about the stabilizing role of generalist species with multiple weak interactions²⁹.

Trophic interactions can display endogenous fluctuations and force us to reconsider the equilibrium nature of the local-regional control of food webs. In large-model food webs, Plitzko and Drossel³⁵ found that intermediate dispersal among food webs limited the occurrence of equilibrium communities but maximized the persistence of species through oscillations and dynamic coexistence. Trophic metacommunities of small food-web modules have explored the local-regional implications of spatially explicit and of oscillatory dynamics, focusing on the maintenance of regional heterogeneity from the interplay between local oscillations and movement. This approach has a long history in both experimental³⁹ and modelling^6,7 studies. The use of synchrony across species and habitats has recently emerged as a powerful tool to understand and predict the dynamics and persistence of spatially extended simple food-web modules^40,41. Gouhier et al.⁴⁰ showed that dispersal and correlated environmental fluctuations interact to affect metacommunity stability through their impact on both interspecific (compensatory dynamics) and intra-specific (spatial) synchrony. Using a similar approach, Pedersen et al.³¹ considered heterogeneous dispersal across trophic levels and showed that tri-trophic food webs are stabilized by the formation of patterns in the presence of both weak and strong dispersal across trophic levels (Figure 1B). These results echo the importance of weak and strong trophic species interactions for food-web stability⁴² and suggest the possibility for an integrated treatment of spatial and species interactions²⁶.

Both competitive and trophic frameworks are readily applicable to evolutionary ecology where both sorting and selection interact with dispersal to drive the assembly of species traits in relation to their environment^14,43 and where body size can be used as a proxy for trophic position^19,44,45. Moreover, phylogenetic approaches to metacommunities can identify the contribution of biogeographical history to both local and regional distribution of species⁴⁶, can more specifically predict increasing phylogenetic structure with community size and isolation⁴⁷, and have improved the testability of metacommunity predictions of patterns of beta diversity⁴⁸. Phylogenetic analysis of metacommunities also suggested that non-equilibrium neutral dynamics of total abundance are more compatible with observed phylogenies than steady-state communities⁴⁹. However, the ‘network of networks’ approach of trophic metacommunities assumes a completely open food-web network, where organic matter is lost from the metacommunity and where inorganic nutrients flow through each local habitat independently. Yet the cycling of (in)organic matter is key to understanding food-web stability⁵⁰ and essential to extending local-regional theories to a broad range of ecosystem functions. This understanding is a long-standing goal of ecosystem ecology and more recently has been addressed through meta-ecosystem theory.