Potential distribution of the oncilla Leopardus tigrinus (Carnivora: Felidae) based on highresolution spatial data and camera traps in a national park in Peru

Abstract

Potential species distribution models (SDMs) are typically evaluated over large territories via low-resolution spatial data and biased presence records. The Parque Nacional de Tingo María (National Park; PNTM, by their initials in Spanish), a small territory, presents challenges for this type of study. This study aimed to analyze the SDM of an oncilla (Leopardus tigrinus) and evaluate the most influential spatial data and specific distribution areas. The Maxent (Maximum Entropy) algorithm was used with 32 high-resolution spatial data layers derived from a PlanetScope multispectral satellite image, DEM, and bioclimatic, social, and infrastructure variables, resampled to 4.5 m resolution, along with 13 presence records from camera traps. The zones were overlaid to identify the species' habitat. The model has an AUC of 0.992, covering 54 hectares, favoring areas with the coldest quarter precipitation (Bio 19), dense vegetation (NDVI), and avoiding roadways and areas with lower elevations. The species is located in Buffer Zones (72.06%), Strict Protection Zones (3.93%), Recovery Zones (3.41%), Special Use Zones (0.54%), Tourist Zone I (10.66%), Tourist Zone II (0%), and Wilderness Zone (9.4%). The viability of the SDM in small spaces with precise, high-resolution data is highlighted. The current zoning of PNTM may not be sufficient to protect the critical habitats of most oncilla species.