Página Inicial » Inscrições Científicas » Trabalhos
Dados do Trabalho
A FORGOTTEN MAMMAL: THE BRAZILIAN FOREST RABBIT (SYLVILAGUS BRASILIENSIS: LAGOMORPHA, LEPORIDAE) IN AGRICULTURAL LANDSCAPES OF SOUTHEASTERN BRAZIL
The Brazilian forest rabbit, also known as tapiti (Sylvilagus brasiliensis), is the single native leporid of South America. Currently, the species is considered the most widespread lagomorph of America, occurring from southern Mexico, along its eastern coast, to northern Argentina. Due to its elusive behavior and nocturnal habit, the ecology of the tapiti remains very little studied. Although tapiti are often considered a forest dweller species, apparently this rabbit has never been seen deep inside pristine forests. Thus, we do not know if it prefers or avoids forest edges. We also do not know yet if the tapiti uses anthropogenic land covers, including pastures. Thus, information on what actually constitutes habitat for the tapiti, especially in agricultural landscapes, is missing. We addressed this information gap investigating the tapiti occurrence in the northeast of São Paulo state, southeastern Brazil. We obtained occurrence data on this species in 205 sampling sites randomly located inside (101 sites) and outside (104 sites) protected areas, during the dry season of 2013 and 2014. We sampled each site with one camera-trap during 30 days. Additionally, we searched for lagomorph tracks two times, during camera set up and removal, respectively, in a 200-m long transect (uncertain method). We assessed several native and agricultural habitat covariates, including forest edges, as predictors of occupancy (ψ) using single season single species occupancy models accounting for false positive detections. The model selection was based on Akaike Information Criterion, corrected for small samples (AICc). We found strong evidence of false positive detection errors on tracks identified as from tapiti. If not accounted for, these errors would inflate in 100% the estimated ψ. We also found that cover by native forests (NaF) and distance from our sampling sites to the nearest edification (Edif) i.e., farmhouses and villages, best explained ψ, both with a positive and well-estimated effect ( NatF = 0.03, SE = 0.008; CI = 0.01 to 0.05 and Edist = 0.0008, SE = 0.00023; CI = 0.00036 to 0.001, respectively). On the other hand, all other assessed potential predictors of ψ, including pasture and edge density, received essentially no support. This study demonstrates that tapiti tracks is highly prone to misidentification, leading to false positive detections that can, if not corrected, substantially bias occupancy estimations. The positive effect of native forests on tapiti occurrence corroborates the hypothesis that this rabbit is indeed a forest dweller instead of being fond for farmland or forest edges. The lower occupancy close to farmhouses and villages is probably explained by the high intensity of anthropogenic disturbance posed by human proximity, particularly a higher incidence of hunting and/or higher predation risk and harassment by free-ranging dogs. Our study shows that native forests are a key habitat for the occurrence of tapiti, corroborating the importance of adherence of rural owners to the Brazilian Forest Act. We also conclude that detecting this species through tracks is prone to errors that can substantially bias occupancy estimations.
tapiti; occupancy modeling; false positive; forest dweller species; anthropogenic disturbance.
This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP 2011/22449-4). We would like to thank Dr. Aurelio Fontes for its support with Geographic Information Systems and mapping; the Instituto Geográfico Cartográfico do Estado de São Paulo (IGC) for providing orthophotomosaic images; the University of São Paulo, the International Paper Co. of Brazil, the Instituto Florestal, and the Fundação Florestal for logistical support and CNPq, CAPES and FAPESP for providing scholarship to NP during his Master and PhD (130198/2014-5, 1772007 and 2018/11788-1, respectively).
Nielson Pasqualotto, Danilo Boscolo, Adriano Garcia Chiarello