Avaliando as plataformas de ciência cidadã para o estudo da migração altitudinal de aves

Abstract

Altitudinal migration is the seasonal movement of individuals from reproductive to nonreproductive areas, and vice versa, along an altitudinal gradient. Studies on the subject are still scarce and, as the costs and logistics of field studies are some of the restraints, using citizen science can be a strategy to ensure advances in understanding this behavior in birds. Thus, we sought to evaluate the efficiency and feasibility of data from the eBird and WikiAves platforms to detect the altitudinal migration of birds in Brazil, also investigating the relationships of behavior with environmental variables. We selected 12 species recognized as migratory by the literature, in addition to six residents for comparison. We defined the Brazilian East as the sampling area and, due to differences in the platforms, for the eBird data, we divided the area into cells and calculated the variables for each one. For WikiAves, we used the municipalities as the sampling unit. First, we ran two zero-inflated binomial generalized linear mixed models (GLMMs) from the eBird data, one for migratory species and one for residents, which was not possible for WikiAves due to the low sample size. Due to the large variance of the species as a random effect factor, we ran new zero-inflated GLMMs for each species using both platforms. The models showed differences in the temporal pattern of occurrence between resident and migratory altitudinal species, confirming the potential of eBird data to infer the species' altitudinal migration behavior. In the species-level models, the resident species data followed the expected, while three migratory ones followed the expected pattern of altitudinal seasonality. In this sense, our models show that the seasonal pattern of occurrence of migrant species was different from the resident ones, as expected in the case of validity of the eBird platform data. The specific models, in turn, confirmed the diversity of behaviors, with only three models of migratory species following the expected one. The diversity of responses, in itself, is plausible and expected for the altitudinal migratory movement, since it is guided by diverse variables that interact with each other and may vary between populations of the same species. In addition to the complexity of the movement, it is also possible that variations between populations result in additional and uncontrollable sources of noise in the models. Even so, the results obtained in the general models proved to be reliable given the established predictions, demonstrating that, in certain circumstances, it is possible to use citizen science data in the study of altitudinal migration. However, more refined analyses at the species level may suffer from noise, possibly making the use of these data feasible only in conjunction or with more refined studies of movement ecology. Thus, we conclude that citizen science data may allow studies of the altitudinal migration of birds in Brazil in certain situations. It is up to future ecology movement studies to refine the scenarios so that citizen science data is more effective.