Introduction: Roads can be a significant barrier to large mammal movements. Roads cause both stress and mortalities for ungulates and in some cases pose a complete barrier - truncating migration corridors and access to seasonal ranges. This is especially true for ungulates, such as mule deer, which have long migration routes that cross multiple roads and overwinter in valley bottoms that often have roads running through them. Although higher traffic volumes are generally thought to increase the barrier effect of roads, there are few guidelines about what traffic volume constitutes a challenge to ungulate movements.
Methods: To address this, we installed thermal video recording systems on eight stretches of rural two-lane highway in Wyoming and captured 1,357 mule deer (Odocoileus hemionus) attempts to cross these highways. For each deer attempt to cross, we measured the duration of gaps between consecutive vehicles that the deer experienced and scored each gap-attempt as successful, failed, or risky (collisions or narrowly avoided collisions). We used logistic regressions to ask whether there were differences in the duration of gaps between consecutive vehicles associated with successful vs. failed and failed vs. risky crossings. Additionally, we used automated radar recorders to measure broader patterns of traffic volume and gaps between consecutive vehicles during >20 nights at each site. From these data, we were able to relate traffic volume per hour with the percentage of the hour that had gaps between vehicles conducive to safe versus failed/risky deer road crossing conditions.
Results: We found strong threshold dynamics in gap duration and traffic volume between failed and safe deer road crossings but not between failed and risky crossings. Patterns from multiple sites indicate that deer need approximately 60 seconds between consecutive vehicles in order to safely cross roads >90% of the time, and a minimum of 30 seconds between vehicles to sometimes cross roads (with risks to both drivers and animals). From the radar recorder traffic data set, we found that a traffic count of 120 vehicles is associated with 50% of the hour having gaps >60 seconds.
Discussion: We conclude that traffic >120 vehicles per hour creates conditions that are difficult for deer to cross roads safely, and traffic >250 vehicles per hour creates an effective barrier to deer movements. By combining this with Wyoming DOT data on spatio-temporal patterns in traffic, we show which roads in Wyoming present the greatest barrier to habitat connectivity for deer. In doing so, we show where highway crossing structures are most needed to ensure mule deer habitat connectivity. This approach sheds new light on how to reduce the barrier effects of roads most cost-effectively. Results can be applied to mule deer throughout the Western USA, and this approach can be used for other species around the world.