Identifying Deer-Vehicle Collision Concentrations in Minnesota

Project summary:

Deer-vehicle collisions (DVCs) are a significant risk to public safety on Minnesota roads, causing property damage, human injuries, and deaths, and also killing deer. In 2019, there were 1,573 DVCs reported to the Minnesota Department of Public Safety (MnDPS), or roughly 4.5 DVCs per day.

Reducing the number of DVCs in a cost-efficient manner would require an analysis of DVCs, using geographic, road type, land use, deer, traffic volume, and other data to identify locations where safety measures or warnings would be most beneficial. In this project, researchers will analyze about 10,000 DVCs to identify factors that increase the risk of a DVC in Minnesota.

Many DVCs without significant vehicle damage or passenger injury go unreported to MnDPS. Researchers will collect roadkill data to estimate the number of unreported DVCs in 2021-2022 in the Duluth area, with the possibility of applying this method more broadly across Minnesota. This data is critical to establish that DVCs which have caused significant property damage or human injury are not different from DVCs that are unreported.

Researchers will use the results of the DVC analysis to construct a data-driven machine-learning-based model to predict hotspots for DVCs. This model will identify road and habitat features that are associated with higher DVC rates.

The products that will be available from this project include:

1. Literature review on DVCs.

2. An interactive map that identifies DVC hotspots.

3. Recommendations for locations where DVC risk could be reduced with highest return on investment for improvements.

4. Pilot project to estimate unreported DVCs in and near Duluth, MN, and development of a protocol to apply across the state.

These products will meet the goal of this project: identify ways to reduce the impact of DVCs on drivers in Minnesota.