Risk Mapping Wildlife-Vehicle Collisions in the State of Montana

Road ecologists and transportation engineers have been exploring new methods to adapt to the environmental and motorist safety concerns involving wildlife-vehicle collisions. There are over one-million crashes with large-bodied wildlife every year in the U.S. that result in substantial property damage and personal injuries. Recent studies modeling these collisions identify where they cluster, and the landscape, road, and driver characteristics that influence the likelihood of a collision along short road corridors and relatively small geographic regions. A cheap and effective method to model wildlife-vehicle collision risk on a large geographic scale is warranted. This research expands on current knowledge and models the risk of wildlife-vehicle collisions in the state of Montana. The large area, diverse landscape, and numerous large mammals that inhabit it make it the perfect candidate.
By dividing the road network into short sections, buffer analysis units are built to extract prediction variables obtained from state and federal open-sourced data websites. Different machine learning methods are used for variable selection and validation. Using an 80/20 split of data for a training and validation set, regularization is used for feature selection to reduce over- and under-fitting a collision risk model. Random forest and extreme-gradient boosting machine learning helps better understand the percent of variance explained by the variables selected. These methods and predictions are compared to the accuracy of negative binomial regression models to evaluate performance. Furthermore, the application of synthetic minority over-sampling of the response variable has the ability to further increase the accuracy of prediction models related to wildlife-vehicle collisions. A range of statistical measurements are taken to compare the models.
The best performing model is applied to a full set of data along Montana's road network to estimate the changes in wildlife-vehicle collision risk throughout the state. Yearlong and seasonal risk is mapped onto the road network and color-coded to show the differences across Montana. The maps capture the changes in risk throughout the year, they match where wildlife-vehicle collisions actually happen, and even coincides with published work on the locations of collision hotspots in Montana.
This research is the basis for future complex real-time risk-mapping models that can be integrated into smart technology and developed into on-board driver alert systems. With the advancements of autonomous vehicles, it is possible to incorporate real-time driving data into models that will analyze wildlife-vehicle collision risk based on vehicle location, season, time of day and driving habits. This can increase driver safety by informing them when they are traveling in areas where wildlife-vehicle collisions are more likely to happen, and can be especially helpful while driving on unfamiliar roads.