Dados do Trabalho

TÍTULO

LANDSCAPE OF FEAR OF THE ANDEAN WHITE-EARED OPOSSUM DIDELPHIS PERNIGRA (MAMMALIA: DIDELPHIMORPHIA) IN A SUBURBAN AREA

Resumo

Deforestation, fragmentation and urbanization of natural areas in Colombian Andes have led to the loss and modification of habitat for native mammals, especially scansorial and arboreal ones. Understanding the impacts of these anthropogenic interventions on Andean mammals’ species, may allow finding alternatives to mitigate their negative effects. The Andean White-eared Opossum, Didelphis pernigra, is a widely distributed marsupial in the Colombian Andes, with a moderate tolerance to human presence. This opossum may move through disturbed areas, thus helping to preserve ecosystem functions in rural and urban environments. We built maps with the geographical representations of space used by the opossum in a suburban area to understand the impact of human activities on the fear perception for those animals, i.e., landscapes of fear. To do so, we used the giving-up densities (GUD) technique, which is the amount of food left by a forager after exploiting a food patch, and allows to evaluate the perceived costs and benefits of foraging. We placed 45 foraging stations at a university campus (75, 5 ha) in the Andes of Colombia (2550 m s.n.m.). The feeders contained 150 mL of a sugar-water mixture (3:1) and 60 glass spheres, to generate diminishing returns in the harvest rate, which is a requirement to measure GUD. We did three samplings: two of 20 nights in 2014 and one of 20 nights in 2015, each sampling after 10 days of habituation period. During the sampling, we measured the amount of food left by the opossums after every night of foraging. The campus is mainly covered by pastures, but it also includes buildings, paved roads, native forest remnants, Eucaliptus plantations, and crops. We built models using the “dredge” function from Mumln Package in R to explain GUD from each foraging station based on landscape features (percentage of coverture by roads, buildings, construction zones, pastures, water bodies and forests),climatic variables (temperature, relative humidity), and percentage of lunar illumination. We used those models to calculate the importance of the predictors with the function “importance” of the same R package.  Then using the Akaike information criterial to choose the best models. The predictors included in the best model to explain variation in GUD were forest and water bodies. The opossums preferred to forage in areas with highest percentage of forest cover and water availability at the study site suggesting that these variables allow animals to feel more secure when exploiting a food patch. Constructions was the least important variable and buildings was not included in any model, contradicting the hypotheses of previous authors that fear is associated to activities at construction sites in the landscape or to human presence. In conclusion, our results indicates that, although D. pernigra has a moderate degree of tolerance to anthropogenic intervention, this mammal feels safer foraging in places with greater forest cover and closer to water. This highlights the importance of the conservation of these type of landscape elements in highly disturbed environments to conserve the ecosystem functions associated to the opossum.

Palavras-chave

 deforestation, giving-up densities, human disturbance, optimal patch use, risk of predation.