Non-shared dispersal networks with heterogeneity promote species coexistence in hierarchical competitive metacommunities

The competition-colonization trade-off has been a classic paradigm to understand the maintenance of biodiversity in natural ecosystems. However, species-specific dispersal heterogeneities are not well integrated into our general understanding of how spatial coexistence emerges between competitors. Combining both network and metapopulation approaches, we construct a spatially explicit, patch-occupancy dynamic model for communities with hierarchically preemptive competition, to explore species coexistence in shared vs. non-shared dispersal networks with contrasting heterogeneities (including regular, random, exponential and scale-free networks). Our model shows that species with the same demography (i.e. identical colonization and extinction rates) cannot coexist stably in shared networks (i.e. the same dispersal pathways), regardless of dispersal heterogeneity. In contrast, increasing dispersal heterogeneity (even at very low levels of heterogeneity) in non-shared networks can greatly promote spatial coexistence, owing to the segregation-aggregation mechanism by which each species is restricted to self-organized clusters with a core of the most connected patches. However, these competitive patterns are largely mediated by species life-history attributes, for example, a unimodal biodiversity response to an increase of species dispersal rate emerges in non-shared heterogeneous networks, with species richness peaking at intermediate dispersal levels. Interestingly, increasing network size can foster species coexistence, leading to a monotonic increase in species-area curves. This strongly suggests that, unexpectedly, many more species can co-occur than the number of limiting resources. Overall, this modelling study, filling the gap between network structure and spatial competition, provides new insights into the coexistence mechanisms of spatial heterogeneity.