Spread rates on fragmented landscapes: the interacting roles of demography, dispersal and habitat availability

Aim We still lack a comprehensive understanding of the relative importance of demographic, dispersal and landscape characteristics on species’ rates of range expansion (RRE) and on how these factors interact. Here, we provide an analysis of these effects for passive dispersers, by investigating how habitat characteristics, such as habitat quality, availability and fragmentation, interplay with species’ dispersal characteristics in determining species’ RRE. In addition, we assessed the predictability of RRE in cases where we have the knowledge of a species’ demography, dispersal and habitat availability. Methods Using the newly available individual-based modelling platform, RangeShifter we simulated the range expansion of species with different dispersal abilities, by varying mean dispersal distance and number of emigrants, on various landscapes. Landscapes varied in habitat quality (in terms of carrying capacity and species’ growth rates) and in habitat availability (in terms of the proportion of suitable habitat and its degree of fragmentation). Results Our results show that 55% of the total variation in RRE was explained by our six main effects, being considerably faster in landscapes with more suitable habitat, but only slightly affected by the degree of habitat fragmentation. Also, synergies between the amount of suitable habitat and species dispersal characteristics had significant positive effects on range expansion. Notably, however, 33% of variation in RRE was not explained by any of the tested factors or interactions between them and can be considered inherent and irreducible uncertainty. Main conclusions Simulation-based approaches provide important insights into the drivers of RRE that are relevant for conservation planning. For instance, our results indicate when it is likely to be better to allocate resources to improve existing habitat rather than creating new habitat, and vice versa. Additionally, our results emphasize that there will often be substantial uncertainty in the RRE, which needs to be taken into account for ecological management.