Local diversity in heterogeneous landscapes: Quantitative assessment with a height-structured forest metacommunity model

"Mass effects," in which "sink populations" of locally inferior competitors are maintained by dispersal from "source populations" elsewhere in the landscape, are thought to play an important role in maintaining plant diversity. However, due to the complexity of most quasi-realistic forest models, there is little theoretical understanding of the strength of mass effects in forests. Here, we develop a metacommunity version of a mathematically and computationally tractable height-structured forest model, the Perfect Plasticity Approximation, to quantify the strength of mass effects (i.e., the degree of mixing of locally dominant and subordinate species) in heterogeneous landscapes comprising different patch types (e.g., soil types). For realistic levels of inter-patch dispersal, mass effects are weak at equilibrium (i.e., in the absence of disturbance), even in some cases where differences in growth, mortality, and fecundity rates between locally dominant and subordinate species are too small to be reliably detected from field data. However, patch-scale transient dynamics are slow following catastrophic disturbance (in which post-disturbance initial abundances are determined exclusively by immigration) so that at any given time, subordinate species are present in appreciable numbers in most patches. Less severe disturbance regimes, in which some seeds or individuals survive the disturbance, should result in faster transient dynamics (i.e., faster approach to the low-diversity equilibrium). Our results suggest that in order for mass effects to play an important role in tree coexistence, niche differences must be strong enough to prevent neutral drift, yet too weak to be reliably detected from field data.