Abstract
Abstract The Ratapani–Kheoni–Singhori landscape of Central India represents an emerging tiger conservation unit connecting multiple protected areas within a rapidly transforming human-dominated matrix. Understanding the ecological processes shaping habitat suitability across this fragmented dry deciduous system is essential for long-term tiger persistence. We applied an ensemble species distribution modelling framework integrating five algorithms (GLM, Random Forest, Support Vector Machine, Boosted Regression Trees and MaxEnt) to predict landscape-scale habitat suitability for the Bengal tiger (Panthera tigris tigris). Tiger presence data were derived from intensive sign-based field surveys (2018–2019), and 22 environmental predictors representing topography, vegetation, hydrology, climate, prey availability and anthropogenic pressure were compiled. Multicollinearity was removed using variance inflation factor screening, and 1,000 pseudo-absence points were generated. All models performed well (AUC = 0.89–0.93), with Random Forest achieving the highest accuracy. Individual models converged on similar spatial patterns, indicating stable ecological signals. The ensemble prediction identified approximately 2006.20 km² of suitable tiger habitat, concentrated within Ratapani Tiger Reserve, Singhori Wildlife Sanctuary and the central forest corridor linking these units, while low suitability coincided with agricultural and human-dominated patches. Variable importance analysis revealed that prey availability, proximity to perennial water sources and gradients of human disturbance jointly structure tiger habitat across this landscape. These results demonstrate how interacting ecological gradients govern habitat suitability and connectivity, revealing the Ratapani–Singhori–Kheoni complex as a single functional ecological unit and providing a spatial foundation for corridor strengthening and conservation planning in Central India.