Understanding the mechanisms that affect habitat use by vertebrates is critical for understanding how species are distributed across landscapes and how they cope with habitat change. The Brazilian Savanna (the Cerrado) has vegetation ranging from grassland to woodland savannas and harbors a rich and diverse amphibian fauna impacted by accelerated habitat loss. Here, we test the influence of vegetation type (from grassy scrubland to woodland) and distance from breeding sites (ephemeral water bodies) on body size, abundance, and survival of the frog Physalaemus nattereri in a natural metapopulation system of south-central Brazil. We also test whether body size is a significant predictor of population abundance. We found that the abundance of P. nattereri varies according to the mean snout-vent length of each metapopulation (sampling unit), as well as a higher estimated mortality rate in woodlands compared to typical Cerrado. Furthermore, we found no difference in estimated mortality among sampling units located far or close to ephemeral water bodies. Thus, our results highlight variable responses of P. nattereri metapopulations to environmental factors, despite the observed high heterogeneity among sampled habitats and the importance of ephemeral water bodies for reproduction. These findings highlight that land cover and availability of breeding sites might not always interact to explain population persistence of Cerrado frogs. 

Sampling was performed monthly for 10 days, from August 2016 to July 2018, for a total of 24 field trips and 240 days of fieldwork. At each sampled vegetation type (campo cerrado, campo sujo, cerrado sensu stricto, and cerradão), three units of pitfall traps with drift fences were installed. Each unit corresponded to two 40-m trap lines, 60 m apart. Each trap line had four 100-L plastic buckets every 10 m, connected by a plastic drift fence 60 cm tall. Dorsal marks of individual frogs were recorded using digital cameras. Then, we performed a computer-assisted matching method, using the Wild-ID program. This software detects recaptures based on an algorithm of Scale Invariant Feature Transform, which extracts and combines the distinctive features in the image, regardless of size and orientation.

Microsoft Excel

R