Spatial scaling in bed-site selection by roe deer fawns: Implications for mitigating neonatal mortality during mowing

When habitat use by field-dwelling animals coincides in space and time with agricultural practices such as spring mowing of meadows, human-wildlife conflicts can have deadly consequences for wildlife. Roe deer (Capreolus capreolus L.) fawns are particularly vulnerable because they hide in meadows during the rearing phase. Thus, a better understanding of the habitat drivers of bed-site selection is critical to mitigating fawn mortality during mowing. Here, we tease apart the among-field (presumably driven by maternal behaviour) and within-field (driven by fawn behaviour) components of bed-site selection of roe deer during the spring mowing season. We collected over 600 fawn bed sites across an environmentally diverse study region. At the among-field scale, we implemented a used versus available design and employed a two-part statistical model (GAMLSS) to identify habitat characteristics that were linked to either fawn presence (vs. absence) or abundance on a given field. At the within-field scale, we compared habitat characteristics at fawn bed-sites with paired random sites using a conditional logistic regression model. At the among-field scale, fawns were more likely to be present, and were more abundant, in fields within more diverse, rural landscapes, with nearby woodland. Surprisingly, fawns were more often present in fields that were near roads and had lower vegetation productivity. At the within-field scale, however, fawns preferred bed-sites which were further from both roads and woodland, but that provided the best visual cover to minimise predation risk. Our findings revealed substantial and novel scale-dependent differences in the drivers of habitat selection of mothers and fawns, which, together, determine the precise locations of bed-sites between and within meadows. These results may aid wildlife managers in identifying areas where there is a high probability of encountering a roe deer fawn so as to initiate targeted searches prior to mowing and, ultimately, mitigate fawn mowing mortality.