Scale-dependent drivers of habitat suitability in the endemic gundi (Ctenodactylus gundi) across Tunisia and North Africa

Global climate change is increasingly restructuring arid and mountainous ecosystems, placing endemic species with restricted dispersal abilities at disproportionate risk of habitat contraction and fragmentation. Understanding how environmental drivers operate across spatial scales is therefore essential to anticipate distributional shifts and guide conservation planning. We investigated the spatial distribution of the Ctenodactylus gundi , an endemic rupicolous rodent of North Africa, using a dual-scale modeling framework integrating macro-environmental predictors across North Africa and fine-scale environmental variables within Tunisia. A total of 25 climatic, topographic, and edaphic variables were combined with occurrence records derived from field surveys and biodiversity databases. We applied MaxEnt to estimate current habitat suitability and project future distributions under two climate scenarios (SSP2-4.5 and SSP5-8.5). Models showed high predictive performance (AUC > 0.90), indicating strong discriminatory power across scales. At the fine-scale within Tunisia, slope and thermal variability exerted the strongest influence, whereas at the North African scale, annual precipitation (BIO12) and temperature seasonality (BIO4) were the dominant drivers of distribution. Future scenarios predict further habitat contraction and fragmentation, particularly under SSP5-8.5, with increasing dependence on elevational gradients and microrefugia as potential persistence zones. These findings demonstrate that the determinants of habitat suitability are strongly scale-dependent and that microtopographic complexity and functional connectivity play a critical buffering role under climate stress. By integrating multi-scale environmental constraints, our approach highlights the heightened vulnerability of endemic rupicolous taxa to climate-driven disruption of corridor continuity and habitat availability, providing spatially explicit priorities for conservation planning in arid mountain systems.