Context Although ∼3% of white-tailed deer are killed on roads each year, no previous study has tested for an effect of roads on deer abundance. This is difficult to do because road density is generally negatively correlated with deer habitat availability. Aims Our goal was to determine whether roads affect deer abundance. Methods First, we used an existing dataset from Pennsylvania, USA, to determine a range of paved road densities representing a significant range in deer per capita mortality. We then conducted a field study in eastern Ontario, Canada, with sample sites for relative deer abundance selected such that (1) road density in the surrounding landscapes varied over this same range, and (2) there were low correlations across landscapes between road density and deer habitat availability. The latter allowed us to isolate the effects of roads from the effects of habitat on deer abundance. We indexed relative deer abundance using a combination of pellet samples and track counts. Key results Unexpectedly, we observed a positive relationship between relative deer abundance and paved road density. Conclusions We speculate that this positive relationship is due to (1) reduced deer predation and/or perceived predation risk and/or hunting pressure in landscapes with higher road density and/or (2) provision of a resource or service by roads, the benefits of which outweigh the road mortality. Implications We found no evidence that road mortality places deer populations at risk of decline, at least over the range of road density values in our study. Therefore we conclude that road mortality is not a conservation concern for white-tailed deer in ecological contexts similar to our study areas.

Additional Metadata
Keywords deer-, habitat fragmentation, Odocoileus virginianus, Ontario, Pennsylvania, reproductive rate, road mortality, vehicle collisions
Persistent URL
Journal Wildlife Research
Munro, K.G. (Keith G.), Bowman, J. (Jeff), & Fahrig, L. (2012). Effect of paved road density on abundance of white-tailed deer. Wildlife Research, 39(6), 478–487. doi:10.1071/WR11152