Metacommunity

A metacommunity is a spatially distributed network of local ecological communities linked by the dispersal of multiple potentially interacting species across a heterogeneous landscape (Wilson, 1992; Leibold et al., 2004). Just as Richard Levins's (1969) metapopulation theory extended single-species population dynamics from an isolated patch to a network of patches connected by migration, metacommunity theory resolves a foundational debate in community ecology—whether local species diversity (α-diversity), turnover between habitats (β-diversity), and regional richness (γ-diversity) are governed primarily by local deterministic processes (abiotic environmental filtering, resource competition, and predation) or by regional spatial and historical processes (dispersal limitation, source–sink immigration, and neutral ecological drift). By treating local communities as open nodes embedded in a regional dispersal network, the metacommunity framework demonstrates that local and regional processes interact along a continuous gradient of dispersal rate and environmental heterogeneity.

In a landmark 2004 synthesis, Mathew A. Leibold, Marcel Holyoak, Nicolas Mouquet, and colleagues classified all existing spatial community models into four canonical paradigms, distinguished by whether habitat patches are environmentally homogeneous or heterogeneous, whether species differ in their niches or competitive traits, and how rapidly propagules disperse between patches:

Across a heterogeneous landscape of M patches with different resource optima, varying the inter-patch dispersal rate (d or m) from zero to complete panmixia drives a characteristic non-monotonic shift across the paradigms—producing a hump-shaped (unimodal) relationship between dispersal rate and local species richness (α-diversity) while causing between-patch turnover (β-diversity) and regional richness (γ-diversity) to decline at extreme dispersal (Mouquet & Loreau, 2003):

1. Very Low Dispersal (d → 0, Dispersal Limitation): Many species fail to reach isolated patches that match their abiotic niche; local richness (α) is low and stochastic, while turnover between patches (β) is high due to random colonization history (priority effects).
2. Intermediate Dispersal (d ≈ 0.01–0.10, Optimal Species Sorting): Every species reaches its optimal habitat patch and outcompetes maladapted residents; regional γ-diversity is maximized and local communities closely mirror local abiotic conditions.
3. High Dispersal (d ≈ 0.10–0.50, Mass Effects Peak): High immigration from neighboring source patches allows sink populations to persist inside patches where they are inferior competitors; local α-diversity reaches its global maximum at the transition from Species Sorting to Mass Effects!
4. Extreme Dispersal (d → 1, Regional Homogenization): The entire metacommunity functions as a single well-mixed patch. Whichever single species has the highest spatially averaged regional fitness floods every patch across the landscape and drives all other species to regional extinction—collapsing α-, β-, and γ-diversity to 1.

To test which paradigm dominates in nature, ecologists apply canonical redundancy analysis (RDA) variance partitioning (Borcard, Legendre & Drapeau, 1992) to decompose the total variation in a species-by-site community matrix across hundreds of lakes, ponds, streams, or forest plots into four components:

• [E | S] (Pure Environmental Fraction): Variance explained uniquely by local abiotic and biotic habitat variables (pH, nutrients, depth, canopy cover, predators) after controlling for spatial coordinates—the signature of Species Sorting.
• [S | E] (Pure Spatial Fraction): Variance explained uniquely by spatial Principal Coordinates of Neighbor Matrices (PCNM / Moran's Eigenvector Maps, MEM) after controlling for environment—the signature of Dispersal Limitation (Neutral / Patch Dynamics) or Mass Effects.
• [E ∩ S] (Spatially Structured Environment): Shared variance where environmental gradients themselves vary smoothly across geographic space.

Synthesizing 158 empirical metacommunity datasets across terrestrial plants, birds, insects, fish, and freshwater plankton, Karl Cottenie (2005) and subsequent meta-analyses (Soininen, 2014) demonstrated that:

• Pure Species Sorting (44%) and Species Sorting + Mass Effects / Dispersal Limitation (29%) account for over 73% of natural metacommunities, whereas purely neutral or purely patch-dynamic patterns without environmental sorting account for <8%.
• Dispersal Mode Predicts Metacommunity Structure: Passive dispersers transported by wind or water currents (such as bacteria, phytoplankton, and zooplankton in isolated ponds) show strong Species Sorting, whereas active overland dispersers constrained by dendritic river networks (such as stream fishes and aquatic macroinvertebrates in headwater vs. mainstem reaches) exhibit pronounced Mass Effects in confluence hubs and Dispersal Limitation in isolated headwaters (Brown & Swan, 2010).

Because natural communities rarely conform to a single discrete box among Leibold et al.'s (2004) four archetypes—often combining environmental sorting, priority effects, and drift simultaneously—Patrick L. Thompson, Laura Melissa Guzman, Luc De Meester, Mathew A. Leibold, and colleagues (2020) reformulated metacommunity theory around three continuous mechanistic axes embedded in a unified stochastic Lotka–Volterra differential equation:

1. Axis 1 — Density-Independent Abiotic Responses (r_i(z_x(t))): How a species' intrinsic growth rate responds to local abiotic conditions z in patch x at time t (abiotic niche breadth and specialization).
2. Axis 2 — Density-Dependent Biotic Interactions (α_ij): The strength and symmetry of intraspecific vs. interspecific competition, mutualism, or apparent competition—ranging from equal interactions (α_ij = α_ii, Neutrality), to stabilizing niche differentiation (α_ii > α_ij, Coexistence), to destabilizing priority-effect dominance (α_ij > α_ii, Alternative Stable States).
3. Axis 3 — Dispersal (I_i(x, t)): Per-capita emigration, distance-decay kernel, and network connectivity across the landscape graph.

This continuous three-axis space integrates classical island biogeography, Janzen–Connell stabilization, priority effects, and spatial conservation planning—proving that protecting regional biodiversity requires conserving both environmental heterogeneity across habitat patches and landscape connectivity corridors tuned to avoid both isolation-induced extinction and homogenization-induced invasion.

1. Wilson, D. S. (1992). "Complex interactions in metacommunities, with implications for biodiversity and higher levels of selection." Ecology, 73(6), 1984–2000. https://doi.org/10.2307/1941449
2. Mouquet, N., & Loreau, M. (2003). "Community patterns in source-sink metacommunities." The American Naturalist, 162(5), 544–557. https://doi.org/10.1086/378857
3. Leibold, M. A., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J. M., Hoopes, M. F., et al. (2004). "The metacommunity concept: a framework for multi-scale community ecology." Ecology Letters, 7(7), 601–613. https://doi.org/10.1111/j.1461-0248.2004.00608.x
4. Cottenie, K. (2005). "Integrating environmental and spatial processes in ecological community dynamics." Ecology Letters, 8(11), 1175–1182. https://doi.org/10.1111/j.1461-0248.2005.00820.x
5. Thompson, P. L., Guzman, L. M., De Meester, L., Horváth, Z., Ptacnik, R., Vanschoenwinkel, B., Viana, D. S., & Chase, J. M. (2020). "A process-based metacommunity framework linking local and regional scale community ecology." Ecology Letters, 23(9), 1314–1329. https://doi.org/10.1111/ele.13568