Evaluating the feeding habitat of the Asian elephant ( Elephas maximus ) using integrated field surveys and remote sensing

Conservation of Asian elephants in human-modified landscapes requires information on diet, forage availability, landscape change, and movement barriers. In Chunati Wildlife Sanctuary (CWS), Bangladesh, we reanalysed records from 187 dung piles and 791 vegetation plots, multi-temporal Landsat classifications, an exploratory railway-crossing inventory, and CCTV detections at a wildlife overpass. The dung-pile-by-taxon matrix contained 888 detections of 142 plant taxa from 49 families. Dung composition varied among sampled months (permutational multivariate analysis of variance [PERMANOVA], pseudo-F8,178 = 5.24, R2 = 0.191, p = .001) and forest beats (pseudo-F6,180 = 2.64, R2 = 0.081, p = .001), although dispersion differences require cautious interpretation. The 14 focal vegetation taxa also varied among months and beats, but effect sizes were small. Their relative abundance differed from dung-detection rankings, consistent with non-proportional resource use rather than a formal preference test. Across the 8,461.1-ha analytical landscape, mapped forest declined by 376.6 ha from 2000 to 2022. Agricultural producer’s accuracy was 45.3%, so agriculture-fallow change remains uncertain. Sixteen potential railway-crossing locations were catalogued, and 50 CCTV detections confirmed overpass use, without estimating crossing probability or mitigation effectiveness. These findings support standardized monitoring of forage composition, forest change, and railway hotspots.