Why Weak Dispersal and Bigger Landscapes Boost Local Biodiversity

One of ecology’s most persistent puzzles has just received a rigorous theoretical solution. For decades, field ecologists have noticed that species confined to small patches of habitat—connected only loosely by the occasional migrant—somehow manage to coexist locally in greater numbers than theory once predicted. A new study published in PLOS Computational Biology by Frederik De Laender, Andrew Gonzalez, Olivia Bleeckx, Dieter Ebert, and György Barabás now explains why this pattern is not a quirk of particular ecosystems but a mathematical inevitability of competitive metacommunities, and why the effect grows stronger as landscapes get bigger.

To understand why this result matters, it helps to step back and consider how ecologists have traditionally thought about biodiversity in fragmented landscapes. The dominant framework, metacommunity theory, treats a landscape as a collection of local communities linked by the movement of individuals. Within this framework, dispersal has long been recognized as a double-edged force. At low rates, the arrival of new individuals can rescue populations on the brink of local extinction and allow species to persist in patches where environmental conditions would otherwise doom them. At high rates, however, dispersal can homogenize communities, swamp local adaptation, and, in some models, actually undermine coexistence by allowing competitively superior species to flood every patch. The intermediate optimum implied by this tension has been invoked so often that it risks becoming an unexamined assumption. What the new analysis demonstrates is that, for competitive communities in heterogeneous landscapes, the low-dispersal end of this spectrum is not merely one option among many—it is where coexistence is mathematically guaranteed to be promoted, and the strength of that promotion grows predictably with the size of the landscape.

The mathematical heart of the study lies in the spatial extension of Lotka–Volterra competition models, a family of equations that has served as the workhorse of community ecology for nearly a century. In their nonspatial form, these models describe how the growth rate of each species is reduced by its own abundance and by the abundances of competitors, with interaction coefficients capturing the per-capita effect of one species on another. Extending them to a spatial setting means writing a separate equation for each species in each patch, with dispersal terms coupling the equations across patches. Such systems quickly become intractable by exact methods as the number of species and patches grows, which is precisely why the authors turned to analytical approximations. Their approach yields a remarkably compact result: local species persistence can be expressed as a balance between the total density of a species across all patches and its local exclusion rate. In other words, whether a species survives in a given patch depends on how much collective biomass it maintains across the whole landscape versus how strongly the local environment and local competitors push it out.

This reframing has a powerful consequence. It converts a question about a complicated spatial dynamical system into a question about geometry—specifically, about the feasibility domain of the underlying competitive system. The feasibility domain is a concept from nonspatial community ecology: it is the region of parameter space, defined by the species’ intrinsic growth rates and their interaction coefficients, within which all species in a community can maintain positive abundances at equilibrium. A community whose parameters fall deep inside this domain is structurally capable of supporting all its members; one whose parameters fall outside cannot, no matter how the abundances are arranged initially. The authors show that local multispecies coexistence in a heterogeneous landscape is governed by this same structural property. The spatial problem, in effect, inherits its answer from the nonspatial one. This linkage is more than a mathematical curiosity—it provides a bridge between two literatures that have largely developed in parallel, allowing insights about competitive communities in single, well-mixed habitats to be translated into predictions about fragmented landscapes.

The landscape-size effect emerges naturally from this framework. When a landscape contains more patches, the total density term in the persistence balance grows, because a species has more opportunities to maintain individuals somewhere in the network. The local exclusion rate, by contrast, is a property of the patch and its immediate competitive environment. As the number of patches increases, the balance tips further toward persistence, and the range of conditions under which multiple species can coexist locally widens. This explains a pattern that field ecologists have documented repeatedly: larger habitat networks harbor more species per patch, not just more species in total. The mechanism is subtle—larger landscapes do not simply offer more habitat or more niches in this model; they buffer each local community against the stochastic and deterministic forces of exclusion by drawing on a larger regional pool of individuals.

Weak dispersal enters the picture because it is what allows this buffering to operate without destroying the heterogeneity that makes it valuable. When dispersal is weak, each patch retains much of its distinct character, and species can maintain different abundances in different places, exploiting the landscape as a mosaic of refuges and source areas. The total-density-versus-exclusion balance remains favorable because individuals produced in favorable patches can, slowly, subsidize populations in unfavorable ones. When dispersal becomes strong, the spatial structure effectively collapses: all patches converge toward similar compositions, the landscape behaves like a single well-mixed habitat, and the advantages of heterogeneity evaporate. The theory thus identifies weak dispersal not as a marginal condition but as the regime in which spatial coexistence mechanisms are strongest.

An important strength of the study is that the authors did not stop at approximation. They confronted their analytical predictions with numerical simulations of the full spatial Lotka–Volterra systems, exploring how the theory fares across gradients of dispersal rate and interaction strength. The verdict is largely reassuring: the approximations hold across most of the parameter space examined. The theory breaks down only in a narrow corner of parameter space where both dispersal and competitive interactions are very strong. In that regime, the effect of dispersal on coexistence becomes unimodal rather than monotonic—coexistence first improves and then declines as dispersal increases, producing the intermediate optimum familiar from earlier theoretical work. That this breakdown requires the joint presence of strong dispersal and strong competition is informative. It suggests that for the many natural systems where competitive interactions are moderate, the monotonic prediction—more dispersal is not better, but weak dispersal is reliably good—should hold.

The treatment of landscape size in the simulations is equally notable. Across the parameter regimes where an effect could be detected, landscape size retained a positive influence on coexistence without exception. Unlike dispersal, whose effect can flip sign under extreme conditions, the landscape-size effect proved robust. This asymmetry matters for conservation thinking, because it implies that enlarging or connecting habitat networks operates through a mechanism that does not depend on fine-tuned conditions. Adding patches to a network consistently expands the feasibility of local coexistence, whereas manipulating dispersal requires knowing where on the response curve a particular system sits.

The empirical component of the study gives these theoretical claims a grounding in the real world. The authors analyzed long-term data from a natural metacommunity of microcrustaceans, a group whose small size, rapid population dynamics, and occupancy of discrete habitat units make them classic subjects for metacommunity research. Two findings stand out. First, the data revealed strong competition among the species, confirming that the competitive assumptions of the models are appropriate for this system. Second, and more strikingly, the dispersal rate showed no detectable effect on local coexistence, while species co-occurrence increased with local landscape size, exactly as the theory predicts. In a field where dispersal is often assumed to be the dominant driver of local diversity patterns, this null result for dispersal—paired with a positive result for landscape size—is a pointed illustration of the theory’s central claim: in strongly competitive systems, the spatial architecture of the landscape, not the rate of movement through it, is what sustains local richness.

It is worth pausing on what this means for the interpretation of biodiversity patterns more broadly. Local species richness, the number of species coexisting at a single site, is shaped by processes operating at multiple scales: regional speciation and extinction, dispersal limitation, environmental filtering, and local species interactions. The new results do not deny this multiplicity, but they isolate one thread with unusual clarity. For communities governed by competition, the local coexistence of species is not primarily a consequence of how much individuals move, provided movement is weak, but of how much landscape is available for the metacommunity as a whole. This reframing shifts attention from corridors and connectivity—long the centerpiece of spatial conservation planning—toward the sheer extent and number of habitat patches, at least for competitive systems in heterogeneous environments.

The analytical link to the feasibility domain also opens a practical avenue for prediction. Feasibility is a property that can, in principle, be estimated from pairwise interaction coefficients and growth rates, quantities that ecologists can measure in controlled experiments or infer from time-series data. If local coexistence in a landscape is governed by the feasibility domain of the underlying competitive system, then measuring these parameters in one place may allow researchers to anticipate coexistence outcomes elsewhere, without needing to characterize every patch in detail. This is a meaningful step toward making spatial biodiversity theory operational rather than merely descriptive.

There are, of course, boundaries to what the study claims. The models assume competitive interactions as the organizing force, and the authors are careful to note that their conclusions concern competitive systems specifically. Communities structured instead by predation, mutualism, or facilitation may obey different rules, and the balance between total density and local exclusion may take different forms. Similarly, the microcrustacean system, while a valuable test case, represents one kind of natural metacommunity; whether the same landscape-size signature appears in systems with different dispersal modes, patch dynamics, or environmental drivers remains an open empirical question. The theory’s breakdown under jointly strong dispersal and strong competition also marks a region where future work is needed to characterize what actually happens when the approximations fail.

Nevertheless, the convergence of three lines of evidence—analytical approximation, numerical simulation, and long-term field data—gives the central result an unusually firm foundation. Weak dispersal promotes local species richness in heterogeneous competitive metacommunities, this promotion strengthens with landscape size, and both facts follow from a single structural principle linking spatial coexistence to the feasibility of the underlying competitive system. For a question that has persisted in spatial ecology for decades, the answer arrives with rare generality: the pattern ecologists kept observing in fragmented landscapes was never an accident of particular systems, but a predictable consequence of how competition, movement, and landscape geometry interact.

Subject of Research: Spatial coexistence theory explaining how weak dispersal and landscape size promote local biodiversity in competitive metacommunities

Article Title: Weak dispersal and landscape size inevitably promote local biodiversity in heterogeneous metacommunities of competing species

Article References: De Laender, F., Gonzalez, A., Bleeckx, O., Ebert, D., & Barabás, G. (2026). Weak dispersal and landscape size inevitably promote local biodiversity in heterogeneous metacommunities of competing species. PLOS Computational Biology, 22(10), e1014028. https://doi.org/10.1371/journal.pcbi.1014028

Image Credits: AI Generated

DOI: 10.1371/journal.pcbi.1014028

Keywords: metacommunity, biodiversity, dispersal, species coexistence, Lotka-Volterra competition, landscape ecology, feasibility domain, spatial ecology, microcrustaceans, competitive exclusion, habitat patches, theoretical ecology