Do roe deer react to wildlife warning reflectors? A test combining a controlled experiment with field observations

Millions of animals are killed by vehicle collisions each year. As mitigation measures, wildlife warning reflectors have become increasingly popular, although clear evidence for their effectiveness is lacking. A reason for inconclusive results in the literature may be that most previous studies on the effectiveness of wildlife warning reflectors compare animal-vehicle collision rates with and without reflectors, a setting characterised by low event rates and weak experimental control. Animal behaviour can be expected to provide a more direct evidence for a possible effect of reflectors. In this study, we analyse roe deer behaviour in the presence of a blue semicircle reflector, one of the most frequently applied wildlife warning reflectors in Germany and other parts of Europe. Behavioural response to these reflectors (classified as no reaction, vigilance, short-distance flight and long-distance flight) was recorded both under controlled experimental conditions with captive roe deer and for free-ranging roe deer at road sections with traffic occurrence. We used generalised linear mixed models (GLMMs) to test if reflector presence induced threat-related behaviour (vigilance, flight) and movement away from the reflectors. We found no significant evidence that the light stimulus emitted by reflectors was perceived as a threat or induced evasive movement. We conclude that our study provides no evidence that blue semicircle reflectors induce behaviour in roe deer that seems suitable to reduce roe deer-vehicle collisions.