PS 10-92 - Why the moose crossed the road: A spatiotemporal assessment of wildlife-vehicular collisions in New Hampshire

Background/Question/Methods

Wildlife vehicular collisions (WVCs) pose inherent threats to wildlife populations and human safety worldwide. Road density is increasing in response to increasing traffic volume and new roads increase fragmentation of wildlife habitats. Roads create barriers that range in permeability to wildlife movement. For some species, a road is a fully impassible landscape feature which is avoidable. Others may attempt to cross roads with variable degrees of success, ranging from outright failure (mortality) to temporary success (crossing without harm). Mitigating the negative effects of WVCs in the USA largely fall to state departments of transportation with guidance from the Federal Highways Agency and local municipalities, largely driven by concerned citizens. Until now, WVCs were not evaluated in New Hampshire where human population, traffic volume, and road density are increasing. Our objectives were to map, summarize, and analyze records of WVCs; to use the data to recommend locations and strategies for mitigation; and to identify a suite of risk parameters that could be considered during new road construction. To that end, we analyzed statewide WVC records from 2002-2019 to visualize spatiotemporal patterns and to identify potential predictors of WVCs (e.g., adjacent habitat) that could be used to inform mitigation efforts.

Results/Conclusions

There were more than 27,000 reported WVCs between 2002-2019, with an average of ~1,500 WVCs per year in New Hampshire. At least 62% of these WVCs were reported to occur on roads with posted speed greater than 40mph. We found positive relationships between WVCs and local road density, which corresponds to population centers within the state. WVCs occurred on roads of all size and functional class, with the largest proportion (33%) occurring on “local” roads. We identified WVC hotspots and assessed road characteristics within these road segments to make WVC reduction recommendations, including enhancements to existing culvert road crossings. Using a combination of statistical approaches (e.g. GAMs), we identified collision, road, and adjacent wildlife habitat attributes that explained observed variability in WVC occurrence. Together these provide the transportation and wildlife departments with data to help inform future road planning and infrastructure decisions. The need for higher accuracy and finer resolution data are among the most important conclusions from this first statewide assessment. Nearly 25% of the WVCs were not accompanied with reliable spatial data and therefore were removed from spatial and statistical analysis.