Shifts from pulled to pushed range expansions caused by reduction of landscape connectivity

Range expansions are key processes shaping the distribution of species; their ecological and evolutionary dynamics have become especially relevant today, as human influence reshapes ecosystems worldwide. Many attempts to explain and predict range expansions assume, explicitly or implicitly, so-called "pulled" expansion dynamics, in which the low-density edge populations provide most of the "fuel" for the species advance. Some expansions, however, exhibit very different dynamics, with high-density populations behind the front "pushing" the expansion forward. These two types of expansions are predicted to have different effects on e.g. genetic diversity and habitat quality sensitivity. However, empirical studies are lacking due to the challenge of generating reliably pushed vs. pulled expansions in the laboratory, or discriminating them in the field. We here propose that manipulating the degree of connectivity among populations may prove a more generalizable way to create pushed expansions. We demonstrate this with individual-based simulations as well as replicated experimental range expansions (using the parasitoid wasp Trichogramma brassicae as model). By analyzing expansion velocities and neutral genetic diversity, we showed that reducing connectivity led to pushed dynamics. Low connectivity alone, i.e. without density-dependent dispersal, can only lead to "weakly pushed" expansions, where invasion speed conforms to pushed expectations, but the decline in genetic diversity does not. In empirical expansions however, low connectivity may in some cases also lead to adjustments to the dispersal-density function, recreating "classical" pushed expansions. In the current context of habitat loss and fragmentation, we need to better account for this relationship between connectivity and expansion regimes to successfully predict the ecological and evolutionary consequences of range expansions.

Version 4 has been peer-reviewed and recommended by Peer Community In Evolutionary Biology (https://doi.org/10.24072/pci.evolbiol.100118). This version (version 4.1) is identical except one correction to the Discussion: "dispersal was density-independent in Tribolium castaneum metapopulations with low connectivity, density-dependent in metapopulations with high connectivity " in v4 should have been "dispersal was density-dependent in Tribolium castaneum metapopulations with low connectivity, density-independent in metapopulations with high connectivity "