Performance of camera trapping and track counts for surveying large mammals in rainforest remnants

Getting information on terrestrial large mammals is particularly difficult in tropical rainforests and in altered landscapes, since the traditionally used method (line-transect census) presents low efficiency in dense vegetation, and is difficult to standardize among heterogeneous, fragmented areas where the small size of patches restricts the length of transects. Aiming to generate information to guide the choice of field protocols for surveying terrestrial large mammals in heterogeneous rainforest remnants, we compared the performance and the correlation between the results of two alternative techniques (track counts and camera trapping), and of two types of bait, in 24 forest remnants in a fragmented Atlantic forest landscape. Techniques resulted in similar observed and estimated richness and species composition at the study landscape, including medium-sized and nocturnal species usually poorly represented in line-transect censuses. Although camera trapping resulted in a higher recording rate of the most common species (Didelphis aurita) and track counts in higher recording rates of some less common species (e.g. Dasypus novemcinctus), observed richness and recording rates of most species were correlated across the 24 sites between techniques. Conversely, the use of different baits strongly influenced results, indicating the importance of standardizing baits in comparative studies. Our results suggest that the two alternative techniques present similar performance and are suitable for studying the factors affecting the distribution of large mammals in altered rainforest landscapes. The choice of field protocols should then focus on the available resources and infrastructure, and on particularities of the study area.