Predicting spatiotemporal patterns of road mortality for medium-large mammals

Area of Science:

• Wildlife Ecology
• Conservation Biology
• Road Ecology

Background:

• Road mortality is a significant threat to wildlife populations.
• Understanding spatiotemporal roadkill patterns is crucial for effective conservation.
• Previous studies often lack detailed spatial and temporal resolution.

Purpose of the Study:

• To model the spatiotemporal patterns of road mortality for seven medium-large mammal species.
• To identify key environmental and intrinsic factors influencing roadkill occurrence.
• To inform targeted management and mitigation strategies for road-wildlife conflicts.

Main Methods:

• Utilized a two-year roadkill dataset from Mato Grosso do Sul, Brazil.
• Employed statistical models relating roadkill presence-absence to environmental variables (land cover, climate, NDVI) and intrinsic risk factors.
• Incorporated remote sensing and weather station data for comprehensive variable estimation.

Main Results:

• Models explained a small fraction of spatiotemporal patterns (<0.23) but showed reasonable discrimination (AUC ≈ 0.70).
• Intrinsic spatial and temporal roadkill risk were the most significant predictors.
• Land cover, climate, and Normalized Difference Vegetation Index (NDVI) were also important factors.

Conclusions:

• Spatiotemporal roadkill patterns provide valuable data for focused management actions.
• Identifying high-risk road sections and time periods complements permanent mitigation measures.
• The approach offers insights for strategic wildlife monitoring and mitigation planning.