The vulnerability of species to roadkill
As landscapes are increasingly fragmented by road networks, many wildlife species are at high risk of mortality on roads. Larger mammalian species with broad home ranges are thought to encounter roads more often due to increased movements, and are thus more likely to be killed on a roadway. Species with lower reproductive rates may have limited capacity to recover from increased mortality.
These predictions framed research by Trina Rytwinski and Lenore Fahrig at Carleton University in Ottawa, Canada. Their results, published in Ecological Applications found all three factors – body size, home range and especially reproductive rate – to be significant predictors of road mortality.
They tested for relationships between these predicted factors and the slope of the relationship between road density and relative abundance of 13 mammal species in southern Ontario, Canada. Four additional species were eliminated from analyses because detections of those species (eastern chipmunk, eastern cottontail, red-backed vole and black bear) was highly correlated to road density.
Response to road density was increasingly negative as body size and home range increased and reproductive rate decreased. The latter was the strongest predictor of population-level response to roads, indicating additional conservation challenges facing species with limited reproductive capacity. The authors conclude that “…priority should be placed on mitigating road effects on larger mammals with lower reproductive rates.”
The authors noted that almost one quarter of the overall variation was not explained by these three main predictors. So while larger body size and limited reproductive capacity may expose these species to increased risk, other explanations are also required. Behavioural responses to road is likely important including attractants such as food or habitat features.
--by Ian Adams
Rytwinski, T., & Fahrig, L. (2010). Reproductive rate and body size predict road impacts on mammal abundance Ecological Applications DOI: 10.1890/10-0968.1
Reader Comments