Animal behavior is shaped by the ability to identify risks and profitably balance the levels of risks encountered with the payoffs experienced. Anthropogenic disturbances like roads generate novel risks and opportunities that wildlife must accurately perceive and respond to. Basic concepts in predator-prey ecology are often used to understand responses of animals to roads (e.g., increased vigilance, selection for cover in their vicinity). However, prey often display complex behaviors such as modulating space use given varying risks and rewards, and it is unclear if such dynamic balancing is used by animals in the context of road crossings.

We tested whether animals dynamically balance risks and rewards relative to roads using extensive field -based and GPS collar data from elk in Yoho National Park (British Columbia, Canada) where a major highway completely bisects their range during most of the year.

We analyzed elk behavior by combining hidden Markov movement models with a step-selection function framework. Rewards were indexed by a dynamic map of available forage biomass and risks were indexed by road crossings and traffic volumes.

We found that elk generally selected intermediate and high forage biomass and avoided crossing the road. Most of the time, elk modulated their behavior given varying risks and rewards. When crossing the highway compared with not crossing, elk selected for greater forage biomass and this selection was stronger as the number of highway crossings increased. However, with traffic volume, elk only balanced foraging rewards when they crossed a single time during a travel sequence.

Using a road ecology system, we empirically tested an important component of predator-prey ecology – the ability to dynamically modulate behavior in response to varying levels of risks and rewards. Such a test articulates how decision-making processes that consider the spatiotemporal variation in risks and rewards allow animals to successfully and profitably navigate busy roads. Applying well-developed concepts in predator-prey theory helps understand how animals respond to anthropogenic disturbances and anticipate the adaptive capacity for individuals and populations to adjust to rapidly changing environments.

Data have been collected and processed following Poulin et al. (2023) "Dynamic balancing of risks and rewards in a large herbivore: further extending predator-prey concepts to road ecology", published in Journal of Animal Ecology. More information on the archived datasets, code and GIS layers can be found in the README.md file.

Program R is required to open the datasets (which are .rds files), as well as the R Markdown code file (.rmd file). The vegetation layers are .tif layers, which can be open in R and/or in other GIS programs.