MaxEnt Modeling for Predicting Suitable Habitats in the North Caucasus (Russian Part) for Persian Leopard (P. p. ciscaucasica) Based on GPS Data from Collared and Released Animals

The first leopards were released into the wild in the Russian Caucasus in 2016 and 2018 as the first step of the Federal leopard restoration program. Leopards were tagged with GPS-Iridium collars. The obtained GPS-coordinates were used to create and verify a mathematical model of the potential habitats of this species and evaluate the possibility of the way choice made by leopards on the move. The modeling was performed by the maximum entropy method using the MaxEnt program based on the data from six leopards (four males and two females). A variety of basic habitat factors and characteristics are reflected in WorldClim bioclimatic indicators, including the data on surface radar imagery, and indices calculated on the basis of Landsat8 satellite image mosaic. To describe and verify biotopes, the field data collected in the Republics of North Ossetia-Alania, South Ossetia and Kabardino-Balkaria were used. The obtained results confirmed a high) accuracy in determining the predicted points. The largest contribution to the common model of potential leopard habitats is made by 10 main factors from the whole set. That list includes the slope steepness, altitude, several climatic characteristics, bioclimatic indices, and the vegetation index. According to the modeling results, the most preferred habitats for the leopard in the Caucasus are not very steep (up to 30°) slopes with grassy vegetation at altitudes of about 800–1500 m above sea level, independent of exposure, but with the average value of snow indices.