Low audibility of trains may contribute to increased collisions with wildlife

Train collisions with wildlife occur worldwide and might be more likely when animals fail to detect trains early enough to perform effective escape behaviour. Detection could be especially limited where tracks curve around hills, reducing visibility and audibility of approaching trains, but no literature has examined this potential in the context of terrestrial transportation collisions with wildlife. At 10 locations in a mountain park where the railway curved around elevated topography, we measured train audibility (as a ratio of train to background sound during approach) and developed a physical model to simulate train audibilities along 45.6 km of the same track. We compared both measured and simulated values to locations of wildlife–train collisions over a 35 year period. More wildlife collisions occurred at locations where measured train audibilities (averaged between train directions at each location) were lower and for the lowest quartile of simulated audibilities. Hill height appeared to reduce train audibility for approaches around curves, but track curvature did not predict audibility overall. Background noise from adjacent road traffic reduced train audibility, as did high train speeds and down-grade travel. Our results suggest that co-occurrence of lower audibility with other risk factors will make it difficult to predict and mitigate collision risk from audibility alone.