Identifying predictors of giant anteater relative abundance in agroforestry landscapes

Abstract

The northeastern region of São Paulo state retains the most relevant Cerrado fragments in the state, immersed in matrices strongly dominated by intensive agriculture and forestry. Medium and large mammal species occur in the region, some of which have specialized habits and live naturally at low population densities. This is the case of the giant anteater, Myrmecophaga tridactyla, a taxon threatened not only in the state of São Paulo, but nationally and internationally, and that still persists in this region. This study aims to estimate and compare the relative abundance of the giant anteater in 55 landscapes with different degrees of natural and anthropized vegetation cover in the northeastern region of the state of São Paulo. This will be done by identifying the key landscape cover and configuration variables in the study region that have a significant effect on relative abundance. We will also investigate whether the relative abundance in landscapes containing evidence of breeding (females detected with young) is distinct from landscapes without evidence of breeding, assessing the importance of compositional (land use and land cover classes) and configurational variables (distance between fragments, edge density, among other metrics) of these landscapes containing evidence of breeding. The data that will be used have already been collected in a previous FAPESP project (Regular Grant FAPESP number 2016/19106-1), where sampling protocols with camera traps were used. A parameterization of occupancy models will be used in the analysis in which relative abundance (mean number of individuals per sampling site), our parameter of interest », varies as a function of these predictors. This parameterization does not require individual identification and models false absences (false negatives), as do most common occupancy models, accommodating variation in the number of individuals available for detection. All analyses will be done via PRESENCE, MARK and in the R package unmarked. The results will indicate the variables related to compositional and configurational heterogeneity of landscapes that are most important to the retention of populations of the species, improving understanding of its conservation status, and providing data for decision making on the best strategies for its conservation. (AU)