Do turtle roadkill hotspots shift from year to year?

Abstract

Freshwater turtles face many threats but roadkill is one of the most serious for many species. Roadkill of turtles is not uniformly distributed across roads but aggregated in certain areas, termed hotspots. A key question in identifying hotspots is whether they are fixed locations or if they shift from year to year because of changes in movement patterns. We compared how one, two, and three years of road survey data compared with the pooled data from four years of surveys. We found 254 turtles during 73 surveys during four years along a 15.5 km road section in Ottawa, Ontario, Canada. The four years of pooled data produced four hotspots (“pooled hotspots”) while each year or combination of years produced from three to five hotspots, four of which approximately corresponded to the pooled hotspots. The average percentage overlap of hotspots between one, two, or three years of survey data and the pooled hotspots ranged from 58.7% to 88.9%. Just one year of surveys sometimes missed one of the pooled hotspots, underestimated the spatial extent of the pooled hotspots, and also sometimes produced an additional “temporary” hotspot. Two years of surveys generally produced better approximations of the pooled hotspots and better identified the spatial extent of those hotspots.