In tropical forests, a domain heavily used and historically destroyed by the human population, frugivorous primates make an important contribution to its regeneration by defecating viable seeds. Their role as seed dispersers is inherently dependent on their movement patterns that directly affect the spatial distribution of dispersed seeds. Many factors, such as fragment size and shape, resource distribution and interaction with conspecific groups can shape daily primate trajectories. However, understanding the effect of these factors on primate movement patterns and their consequences on seed dispersal is a challenging task, especially when considering the efforts needed in the field and the numerous confounding factors. The development of mechanistic models such as the agent-based models (ABMs) provides an opportunity to answer this question with reduced field work and in controlled conditions. In this sense, the main objective of this project is to develop an ABM able to simulate the seed dispersal spatial result (or the seed shadow) of an endangered primate, the black lion tamarin (BLT), in different patches of forests with varying characteristics (size, shape and resource distribution). To this end, we are extending an ABM that has been partly adapted to the model species but of which adaptations were not sufficient to reproduce satisfactorily the patterns of movement and seed dispersal observed in nature. Specifically, simulated daily distances were shorter and re-visitation of fruiting trees were more frequent than observed in nature, resulting in half of daily travelled distances and seed dispersal distances. To obtain a model able to simulate real world BLT movement patterns, we will implement new specific rules, that we are calling model extensions, following the Pattern-Oriented Modelling approach. These extensions are related to 1) Step length and turning angles; 2) Resource visitation rules and 3) Territoriality. The first extension is implemented with simple code, and we already have empirical data to parameterize it. The second extension might demand the inclusion of a combination of factors, such as tree size and tree clumping degree. The last extension will be done with mechanistic rules of home range occupation, such as the time needed to visit the entire home range. The last two implementations will require empirical data analysis and thoroughly coding. Results will foster the understanding of how anthropogenic activities such as fragmentation and habitat loss impairs forest regeneration though indirect effects on the seed dispersal fauna.
News published in Agência FAPESP Newsletter about the scholarship: