Monitoring wildlife passages allow scientists to improve effectiveness of these structures by adapting their conception, location and planning. To perform the surveys, infrared triggered cameras are probably the most used technique. Indeed, these small photographic cameras with an invisible flash are well adapted to capture a wide variety of species, even when used inside the narrowest underpass.However, few studies have investigated the efficiency of these devices to systematically and automatically capture a majority of the passage events (PE), or their impact on wildlife passages frequentation. How many animals are missed by the device? Is the reliability dependent on type of species ? What are the best sensors’ parameters? Is there an impact of the devices on animals’ behavior? How many days are needed to make a sufficient survey of an underpass? These are the questions we asked to improve the knowledge in triggered camera utilization.From April 2012 to June 2013, we monitored 12 underpasses located in Alsace (East France) with both triggered photographic cameras (Reconyx HC600) and video cameras. Video cameras recorded 24 hours a day and 7 days per week in order to observe all PE of all species and identify every event missed by triggered cameras configured in 5 different motion/thermal ratios. The behavior of carnivores was recorded in order to assess possible disturbance by the devices on the animals. Finally, we performed a correlation between the number of monitoring days and the diversity of species observed.We found that in small box culverts, 47% of small mammals (voles, mice, shrews) PE and 17% of medium-sized mammals PE (foxes, badgers and other mustelids) are missed by triggered cameras. Moreover, we demonstrated that whatever the season, motion/thermal ratios favoring the motion sensor missed less events than ratios favoring the thermal sensor. Despite every species likely to use our culverts are detected, we found that a minimum of five months were needed to detect 90% of the overall specific diversity over the study period, 3 months more than the duration recommended in previous studies. Monitoring devices disrupted medium-sized mammals during the first five weeks after their installation. There were also fewer carnivores the night after our presence on the field to replace batteries and memory cards.Even by using the latest technologies, our results point out a significant underestimation of wildlife passage utilization which can introduce an important bias to conclusions about effectiveness of faunal passage structures.
wildlife passages ; monitoring ; triggered camera ; small mammals ; culverts ; landscape ecology ; fragmentation