A geo-statistical approach to model Asiatic cheetah, onager, gazelle and wild sheep shared niche and distribution in Turan biosphere reserve-Iran

Abstract

Presence data for four mammals in the Turan Biosphere Reserve in Iran including the Asiatic cheetah ('Acinonyx jubatus venaticus'), the Persian onager ('Equus hemionus onager'), the wild sheep ('Ovis vignei'), and the gazelle ('Gazelle Bennettii') were used to analyze and model their potential interaction, facilitation, habitat coverage and niche dimensions. A geostatistical approach using the spatial autocorrelation between the locality points, and their relationship with habitat resources and characteristics with application of remotely sensed maximum enhanced vegetation index (EVI) and surface temperature, elevation, aspect, vegetation cover and soil moisture was used to predict herbivores species niche. The potential suitable habitat of herbivores along with environmental variables was used to model the predator species (cheetah) niche. The model results were tested using fivefold cross validation by area under the curve (AUC) values on set of independent testing data and were compared to more commonly used models of generalized linear model (GLM) and MaxEnt. The results show that cheetah's potential suitable habitat has 61% overlap with wild sheep, 36% with onager, and 30% with gazelle. Onager habitat has 64% overlap with gazelle and 60% the wild sheep. Wild sheep on the hand, shares only 37% of its habitat with gazelle. The most prey and predator interaction exists between cheetahs and wild sheep, while onagers provides facilitation for gazelles and wild sheep by potentially providing extra water sources. Among the implemented modeling techniques, spatial GLM showed better performance over GLM and MaxEnt. We suggest that conservation effort should focus more on maintaining the population of wild sheep and onagers to support other species in the habitat.