Modeling animal-vehicle collisions using diagonal inflated bivariate Poisson regression

Area of Science:

• Transportation Safety
• Statistical Modeling
• Wildlife Management

Background:

• Discrepancies exist between reported animal-vehicle collision (AVC) data and carcass removal data.
• Existing models struggle to reconcile these two data sources effectively.
• A comprehensive understanding of AVCs requires integrated analysis of available data.

Purpose of the Study:

• To develop and apply a novel statistical model for analyzing discrepancies in AVC data.
• To integrate reported AVC data and carcass removal data for improved risk analysis.
• To provide a better understanding of highway animal-vehicle collisions.

Main Methods:

• Utilized a diagonal inflated bivariate Poisson regression model.
• Applied the model to reported AVC and carcass removal data from Washington State (2002-2006).
• Compared the proposed model against double Poisson, bivariate Poisson, and zero-inflated double Poisson models.

Main Results:

• The diagonal inflated bivariate Poisson model effectively fits correlated and dispersed AVC data.
• The model demonstrated superior performance in fitting datasets with overlapping information.
• Identified key factors influencing AVCs, including traffic volume, speed limits, and road geometry.

Conclusions:

• The diagonal inflated bivariate Poisson model offers a robust approach for analyzing AVC data.
• This method enhances the understanding of factors contributing to animal-vehicle collisions.
• The findings provide valuable insights for implementing targeted highway safety measures.