Talk Making light in amphibian road tunnels; Novel automated technology shows large seasonal life stage and speciesspecific variation in amphibian usage in the UK
Abstract

Road mortality, habitat fragmentation and loss of connectivity represent major threats for amphibian populations worldwide and especially in Europe where the road network is exceptionally dense. Road mitigation using tunnels represents the most promising solution yet most tunnels remain poorly investigated and insufficiently monitored, largely due to methodological constraints and high costs of pitfall trapping. Since 2013 we developed and deployed a custom-made automated system for monitoring such tunnels and investigated multiple sites in the UK and Sweden over three years. We used high-frequency timelapse infrared image recording for relatively cheap and robust monitoring and undertook intensive data collection focusing on the amphibian maximum periods of activity (spring, late summer and autumn). We recorded over 5.3 million images and developed a software script for image analysis which reduced the workload by as much as 85%. We categorised all amphibian observations by species, age class, sex (for some species), time and position inside the tunnel as well as behaviour (e.g. crossing, foraging, turning, hesitant, etc.) and directionality (moving in or out of the tunnel). Since 2014 we investigated the almost completely unknown aspects of amphibian road mitigation usage by post-metamorphic juveniles following emergence from the ponds. We recorded over 13000 amphibian observations from 5 species at the six different sites as well as over 1500 records of other species including 4 species of reptiles and 11 species of mammals. Where possible we used site context and data to interpret tunnel usage at a population level. We used generalised linear models to understand the relationship between amphibian success rate during tunnel crossing, specie, life stage and season. A large percentage of all amphibian observations did not result in a full crossing but rather in/out movement at one tunnel entrance and this percentage varied significantly with species, site and season. There were substantial differences between speed of travel inside the mitigation system by different amphibian species, with common frogs (Rana temporaria) typically the fastest. Juveniles of some amphibian species, especially common frogs and common toads (Bufo bufo), used the tunnels almost exclusively during the day compared to newt species which were entirely nocturnal, suggesting important differences in road traffic impacts on juveniles in unmitigated sites. Overall, road mitigation appeared significantly more successful for some species than others at different sites. Crossing success rates were very poor for some reptile species but not others. Effectiveness in maintaining habitat connectivity needs to be evaluated carefully at species level rather than in large taxonomic groups. Automated passive monitoring systems such as our timelapse camera systems can gather highly detailed data over long periods of time in a cost-effective manner and can transform our knowledge of road mitigation efficacy.

Keywords
Amphibian tunnels, Camera, Mitigation, Effectiveness, Road kill, UK