The dispersal of individual seeds or animals varies within and among populations: e.g. the timing and distance of dispersal can differ widely. Understanding this variation in dispersal would benefit ecological forecasts, and likely requires that researchers take into account what the relevant traits of a species are, what the main differences among individuals are and how individuals respond to environmental cues. A complication is that dispersal data are often of an indirect nature: for instance when researchers ‘only’ have records of the distribution of individuals after dispersal relative to the distribution of potential sources (e.g. seed traps in forests and fields), or the locations of individuals that have not completed dispersal yet (e.g. sightings of marked individuals of migratory species). In all these cases inverse modelling can be a useful tool to go from pattern data to parameter estimations of hypothesized processes and causes. In this talk we will explore this approach in the light of dispersal in three very different systems, which include variation in seed dispersal in tropical forests, migratory dynamics of geese and weather-dependent dispersal in moose.
Results/Conclusions
In order to parameterize population models for tropical trees, we needed size-dependent estimates of the fecundity of adult trees. Inverse modelling based on the seed density in traps and on the location and size of adult trees, resulted in realistic estimates of dispersal kernel and seed production functions for 18 species. In the case of color-banded Greater White-Fronted Geese, we were interested in the dispersal rates within the wintering areas as well as seasonal migration. Doing so we found that on a monthly basis these geese were very mobile, revealing complex spatial structuring, and rendering local management strategies inadequate. Finally we use the case study of moose equipped with GPS collars to show how process-based models can link the onset and cessation of dispersal to weather variables. Overall these examples show that inverse methods are widely applicable in dispersal ecology and the quantification of movement, and their broad application could profoundly increase our understanding of dispersal and its impact on ecosystems.