Habitat preferences of reintroduced banteng (Bos java']javanicus) into the Salakphra Wildlife Sanctuary, Thailand

Context. Large forest-dwelling mammals are highly sensitive to habitat structure. Thus, understanding the responses of reintroduced banteng (Bos javanicus d'Alton 1823) to their habitat is important for ensuring the sustainability of a reintroduction program. Aims. The aim of the present study was to evaluate the habitat preferences of banteng after reintroduction into the Salakphra Wildlife Sanctuary in Thailand on the basis of fieldwork conducted between January 2015 and November 2017. Methods. Seven banteng individuals bred at the Khao Nampu Nature and Wildlife Education Center were systematically reintroduced into the Salakphra Wildlife Sanctuary in 2015 (four individuals) and 2016 (three individuals). The banteng individuals were tracked via radio-collars and camera-traps. The maximum-entropy method (MaxEnt) and multiple logistic regressions (MLR) were used to identify habitat preferences. Kernel-density estimates (KDE) and a minimum convex polygon (MCP) were used to estimate the area of the habitat used. Key results. In total, 407 radio-signal locations showed that the MaxEnt habitat-preference models classified the banteng as associated with distance from villages and salt licks (regularised training gain of >1.0). Multiple logistic regressions form 32 camera-trap locations classified the banteng as associated with low elevations far from villages, guard stations and roads in a flat area (no aspect). The two methods for estimating habitat use provided similar results and showed that the reintroduced banteng used a wider range of habitat in the dry than in the wet season. Conclusions. The results from the present study suggest that the reintroduced banteng individuals prefer low elevations and flat areas without human activity.