Beyond crippling bias: Carcass-location bias in roadkill studies

Abstract

Wildlife roadkill studies need to cope with a mismatch among recorded carcasses and actual road mortality, because of the existence of three biases: crippling, carcass-persistence, and observer bias. Here, we focused on the often overlooked crippling bias, suggesting that it should be called carcass-location bias and disentangling the related three possible outcomes for affected wildlife: injured animal escaping and dying away from road, animal rebounding off the road after vehicle collision, and animal retained by vehicle. Such outcomes can probably be affected by different species traits, and, in order to make a first evaluation of this hypothesis, we opportunistically collected 150 direct observations on the ultimate fate of roadkilled vertebrates. Approximately one third of them were affected by carcass-location bias, so extremely difficult to be recorded through typical roadkill surveys, entailing a considerable and overlooked source of error for roadkill studies and mitigation actions based on them.