Multispecies assessment of core areas and connectivity of desert carnivores in central Iran

Abstract

Aim Central Iran is a priority area for biodiversity conservation, which is threatened by encroachment on core habitats and fragmentation by roads. The goal of this study was to identify core areas and connectivity corridors for a set of desert carnivores by predicting habitat suitability and calculating resistant kernel, factorial least‐cost path modelling and graph network indices. Location Iran. Methods We used an ensemble model (EM) of habitat suitability methods to predict the potential habitats of leopard, cheetah, caracal, wild cat, sand cat and grey wolf and used resistant kernel and factorial least‐cost path modelling to identify important core habitats and corridors between patches. We also used a graph network analysis to quantify the importance of each core patch to landscape connectivity. Results Potential habitats of the studied carnivores appeared to be strongly influenced by prey density, annual precipitation, topographical roughness, shrubland density and anthropogenic factors. Most of the core patches were covered by protected areas and no‐hunting areas. This may be attributed to the relatively high resistance outside protected areas leading to isolated occupied patches. Patch importance to connectivity was significantly correlated with patch extent, density of dispersing individuals and probability of occurrence in the core patch. Main conclusions Our findings revealed that prey abundance in core habitat is critically important, and has higher influence than habitat area per se. In addition, our analysis provided the first map of landscape connectivity for multiple species in Iran and revealed that conserving these species requires integrated landscape‐level management to reduce mortality risk and protect core areas and linkages among them. These results will assist the development of multispecies conservation strategies to protect core areas for carnivores.

Combining ensemble models and connectivity analyses to predict wolf expected dispersal routes through a lowland corridor

“…For the ultimate purpose of this study, the most suitable approach was therefore to consider variables calculated on a fine scale. To avoid multicollinearity among variables, we checked pairwise Pearson's correlation coefficient between covariates and verified that no variable pair had a coefficient higher than 0.60 [1].…”

“…For this study, the ensemble models were implemented using three species distribution models, including generalized linear models (GLM), generalized boosted models (GBM), and a maximum entropy algorithm (MaxEnt). These species distribution models were selected based on their high predictive power [1,40,41,42]. For each model, we ran three replications where 75% of the occurrence points was used as training set, while the remaining 25% was used for model evaluation.…”

“…Large carnivores are top predators playing a key role in the maintenance of healthy and functional ecosystems [1]. Despite their ecological importance, most large carnivores suffered major declines in both population size and geographic range over the twentieth century [2].…”

Accounting for Connectivity Uncertainties in Predicting Roadkills: a Comparative Approach between Path Selection Functions and Habitat Suitability Models

“…Placement of mitigation structures can be optimized by using models that draw on more easily available data, such as SDMs, because such models may adequately reflect the interactions between species and environment across spatial scales (Guisan et al 2013). The SDMs cost-benefit tradeoff is enhanced when we take into account the growing demand for multiple species assessments (Khosravi et al 2018), as the collection of a comprehensive multispecies pathway dataset is hampered by limited budgets. We therefore support such an application, in the context addressed herein, for other species and with different behaviors or life stages, as long as the SDMs modeling framework is explicit and justified within the context of conservation purposes (Guisan et al 2013;Tulloch et al 2016).…”

“…Ecological corridors applied across wide study areas (e.g., regional scales or higher) normally better rely on dispersal resistance surfaces, since when constructed through habitat suitability or movements within the home range, such corridors may greatly overestimate resistance to movement (Blazquez-Cabrera et al 2016;Ziółkowska et al 2016). In such contexts, the use of a negative exponential transformation function (Trainor et al 2013;Keeley et al 2016) may be more relevant as the generally lower resistance to movements renders distant patches more reachable by a few dispersers, thereby increasing the confidence of corridor placement (e.g., Khosravi et al 2018). Lastly, we argue that these corridors might be likely to benefit other dispersing forest mammal species, given the similar dimension of home range to the species we examined, which in turn can be translated into a capability of similar dispersal distances or suchlike (Bowman et al 2002).…”