Anthropogenic disturbances such as habitat loss and the introduction of invasive species have led to the extinction of numerous species throughout the world. Some of these species provided key ecological functions, and their extinctions contribute to ecosystem degradation. Thus, these systems should be of high priority for rewilding to restore ecological function. On Guam, all native avian seed dispersers have been functionally or completely extirpated due to predation by the invasive brown treesnake (Boiga irregularis). This has led to the degradation of native forests throughout the island. Restoring seed dispersal is essential for the prevention of further degradation, and efforts to rewild Micronesian Starlings (Aplonis opaca) through range expansion are being considered, as well as reintroducing the Mariana fruit dove (Ptilinopus roseicapilla) from neighboring islands. To understand the relationships between different rewilding candidates and small-scale landscape patterns, we developed an individual-based model that simulates the behavior of starlings and fruit doves within different theoretical landscapes. We used data collected in the Marianas for the past 5 years to calibrate starling and fruit dove behavior. We then explored the influence of landscape configuration on dispersal patterns to derive management recommendations and ideal rewilding candidates in relation to landscape structure.
Results/Conclusions
In our theoretical landscapes, we simulated both native and degraded forest cells in different configurations. When varying the spatial configuration of these cells, we observed variations in patterns of seed rain. Heterogeneity of native cells highly influenced the dispersal of native seeds, which are considered in management scenarios as high priority. Highly homogeneous landscapes had the lowest native seed dispersal into degraded areas. Seed disperser behavior also played an important role on seed rain, with Micronesian starlings dispersing native and non-native seeds evenly across the landscape and Mariana fruit doves dispersing native seeds into degraded areas at higher densities. We used these results to identify ideal rewilding candidates and inform potential management recommendations depending on landscape configuration. This model can identify optimal strategies for future rewilding projects that aim to restore seed dispersal to degraded landscapes.