Range expansion is a fundamental biological process with important implications for conservation (e.g. invasive species or species responding to climate change). Recently, research has uncovered the important role that rapid evolution plays during range expansion. Differential selection in the range edge compared to the core can lead to changes in fitness and dispersal ability, thus increasing spread rates on average. However, controlled studies have demonstrated that evolution can also drastically increase the intrinsic variability of range expansions through stochasticity in evolutionary trajectories. It is unknown, though, how rapid evolution during range expansions might interact with other processes known to affect variability in expansions. Landscape structure seems especially likely to interact with spatial evolution during range expansions, but this interaction is difficult to fully capture in experiments or simple, analytical models. Thus, we constructed an individual-based model allowing evolution of both individual fitness and dispersal ability. We coupled the model with experimental data on range expansions in microcosms of the red flour beetle (Tribolium castaneum) to examine the interactions between habitat heterogeneity and trait evolution during range expansion.
We parameterized the model to the red flour system using a separate, previously published data set. We then simulated range expansions through heterogeneous and homogeneous landscapes, while varying the strength of evolution of both traits via a heritability parameter. Finally, we compared the model results with independent data from a parallel experiment in the T. castaneum system. Both model and experimental results suggest that habitat heterogeneity dampens the effect of spatial evolution on variability during range expansions. At low levels of heritability in both traits, simulated range expansions proceeded faster and with more variance in heterogeneous landscapes. However, with increased heritability, this pattern reversed and homogeneous landscapes displayed both faster and more variable spread. Examination of the trajectories of trait evolution suggest this discrepancy between landscape types is due to the added cost of dispersal in heterogeneous landscapes (i.e. individuals can experience a fitness cost due to dispersal into a low-quality habitat). This combination of experimental results with a theoretical, mechanistic model provides important insights into the interaction between rapid evolutionary changes and landscape heterogeneity in range expansions.