Large-Scale Model-Based Assessment of Deer-Vehicle Collision Risk

Hothorn, Torsten; Brandl, Roland; Müller, Jörg · 2012 · OpenAlex-citations

DOI: 10.1371/journal.pone.0029510

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of estimating ungulate population densities and managing deer-vehicle collisions (DVCs) on a large scale. Direct methods for counting deer are costly and lack spatial representation, while existing management systems are often limited to small areas. The authors propose using DVC data as a cost-efficient, indirect relative index for deer density, aiming to model collision risk across Bavaria, Germany, to inform wildlife management and road safety planning. The researchers analyzed data from approximately 74,650 reported DVCs in 2006 and 2009 across 2,223 Bavarian municipalities. They combined this collision data with environmental variables, including climate metrics from WorldClim, land-use classifications from Corine 2000, and browsing intensity data derived from surveys of over 3 million tree saplings. The statistical approach utilized a two-step additive Poisson model. First, a parametric model accounted for road type, year, and red deer presence. Second, a nonparametric model estimated deviations based on environmental factors and spatial heterogeneity using component-wise boosting. This method allowed for nonlinear relationships and variable selection, resulting in a "DVC index" that quantifies local collision risk relative to a baseline. The results demonstrated that the model could predict DVC risk with high precision at the municipal level. The DVC index revealed significant spatial variation, with higher risks in productive agricultural areas south of the Danube River and lower risks in the Alps and large urban centers like Munich. Climate variables contributed most significantly to the model, followed by land use and spatial factors. Specifically, higher minimum temperatures and moderate precipitation levels increased collision risk. Land-use analysis showed a nonlinear relationship with forest-edge length and a negative correlation with urban area proportion. Crucially, the study found a positive correlation between the DVC index and both browsing intensity and official deer harvest numbers, validating the index as a reliable surrogate for deer density. The significance of this work lies in providing a scalable, data-driven tool for ungulate management. By establishing that DVCs can accurately reflect deer density and browsing pressure, the authors offer a cost-effective alternative to traditional survey methods. The proposed DVC index can assist in planning road protection measures, such as fencing, and in defining hunting quotas. Furthermore, the open-source software implementation allows this modeling framework to be transferred to other regions and species, offering a generalizable solution for assessing wildlife-vehicle collision risks in complex landscapes.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• incidence prevalence