Factors influencing the timing and frequency of moose-vehicle collisions at urban-wildland interfaces in subarctic Alaska

Abstract

Wildlife-vehicle collisions concern road engineers, wildlife biologists, and the motoring public. In Alaska, moose-vehicle collisions (MVCs) are the most commonly reported type of wildlife-vehicle collisions. Each year an average of 101 MVCs were reported in the Fairbanks North Star Borough (FNSB), resulting in damages amounting to $3,000,000/yr. This thesis describes the spatial and temporal patterns of MVCs in the FNSB and uses these patterns to infer the interactions between human and moose behavior that cause them. The analytical approach used combined spatial and temporal records of MVCs collected by the Alaska Department of Transportation with spatially explicit data describing topography, land cover, traffic volume, and traffic speed. Multiple hypotheses about cause and effect were tested using computer-intensive, randomization procedures. MVCs occur most frequently during the first hours after sunset, particularly in autumn and winter. Roads in the vicinity of areas of recent wildland fires have a heightened risk of MVCs, particularly if there are moderate traffic volumes and speed limits of 90 km/h (55 mph). MVCs are also frequent on roads traversing land cover types where human population densities are low. Risk of MVCs in the FNSB is highest between 150 m and 200 m elevation. Based on these results, several mitigation measures to reduce MVCs in the FNSB are recommended, including seasonal warning signage and speed reductions in the hours after sunset. Roadside fencing designed to divert moose to designated road crossings in conjunction with infrared-triggered warning lights at these crossing points may be warranted in areas identified as hotspot locations for MVCs.