STKDE+ approach reveals wildlife-vehicle collision hotspots at broken fence locations

Road fencing is considered one of the most effective traffic safety measures to avoid wildlife-vehicle collisions (WVC). Construction errors during fence installation, intrinsic defects and other inconsistencies allow, however, for wildlife to enter the roads. WVC thus occur even at many fenced places. We analysed WVC, from fenced D10 and D11 motorways in Czechia, which took place between 2009 and 2019. We applied the spatial-temporal KDE+ method (STKDE+) in order to determine both spatial and temporal WVC hotspot patterns. We further compared the WVC temporal pattern before and after the installation of fences. Three hotspot patterns were described: hotspot disappearance (after proper design and construction of a fence), emergence of a new hotspot (at places with fence gaps or improper fence ends) and stability of a hotspot over time (non-effective fences). We found that improper design of fencing induced WVC occurrence. We further concluded that the STKDE+ method can be used to effectively determine the effectiveness of safety measures. The STKDE+ analysis revealed 32 WVC hotspots, which contained 46% of all WVC. The fence-end effect and fence gaps caused WVC in 20 and 5 cases, respectively.