Modeling habitat suitability for the Persian leopard in an ecological hotspot: Bamu National Park, southern Iran

Farid Shahidinejad¹Ali Frazam¹Parvaneh Sobhani²Gholamreza Ghaderi Mozafari³Azade Deljouei⁴Marina Viorela Marcu⁵Seyed Mohammad Moein Sadeghi^4*

• ¹University of Tehran, Tehran, Iran
• ²Lorestan University, Khorramabad, Iran
• ³Pars Wildlife Guardians Foundation, Shiraz, Iran
• ⁴Northern Arizona University, Flagstaff, United States
• ⁵Universitatea Transilvania din Brasov, Brașov, Romania

The Persian leopard (Panthera pardus saxicolor), the largest felid in the Middle East, is an endangered subspecies persisting in fragmented mountainous habitats across Iran, where it faces escalating threats from habitat degradation, poaching, and human–wildlife conflict. Bamu National Park (BNP), located in the southern Zagros Mountains, serves as one of the species' last strongholds and ecological hotspot in southern Iran, yet its habitat suitability remains poorly quantified. In this study, we used a maximum entropy (MaxEnt) model to identify suitable habitat and the key environmental and anthropogenic drivers shaping the spatial distribution of Persian leopards in BNP. Presence data were derived from 42 verified leopard occurrence records collected between 2015 and 2017. Twelve predictor variables were retained out of an initial set of fifteen after multicollinearity screening, selected based on ecological theory, previous research, and expert consultation. These included topographic factors (slope, aspect, ruggedness), climatic variables (mean annual temperature and precipitation), vegetation and rangeland types, prey availability (Ovis orientalis), and human disturbance (proximity to water troughs, ranger stations, roads, and the oil refinery plant). The MaxEnt model exhibited excellent predictive performance (mean AUC = 0.959; TSS = 0.84; OR = 0.06). Distance to artificial water troughs was the most influential variable, contributing over 50% to the model's explanatory power, followed by vegetation type and rangeland classification. Terrain ruggedness, prey availability, slope, and aspect were also important, confirming the Persian leopard's preference for rugged, shrub-dominated landscapes with reliable prey resources. These results highlight clear conservation priorities within BNP, including the protection and careful management of core habitats surrounding anthropogenic water sources, restricting road expansion in high-suitability zones, and managing rangeland and vegetation types that support key prey populations. Beyond BNP, this study provides a replicable modeling framework to guide conservation of large carnivores in other mountainous and fragmented landscapes where apex predators face similar ecological and anthropogenic constraints.