Estimativas de detecção de primatas e validação de modelos preditivos em duas unidades de conservação na Amazônia, Roraima, Brasil

Abstract

This study compared the detection of common primate species in the Viruá National Park and in Ecological Station of Maracá, the detections with environmental variables and the generalities of predictive models were tested. We collected data from the primate community of 10 km daily by linear transect method in five campaigns of 60 km in each area. Data on tree and palm density, basal area, canopy opening and canopy height were collected in sub samples of 200 m2 every 100 m along the 60 km, that comprises each of the two trail grids of the RAPELD system used by PPBio. The analysis of the detection probability and validation of predictive models were made for Alouatta seniculus, Saimiri sciureus and the congenerous Cebus apella, Cebus olivaceus, Ateles paniscus and Ateles belzebuth. For all species and congenerous the detection probability was not different in both areas, despite the structural differences. The probability of occurrence was different for congenerous of Ateles between areas. The predictive models were not able to predict the species distribution in the neither area. This indicates that the species have different responses to distinct places and that indirect environment variables are not good predictables for generalities of predictive models.