Journal of Interdisciplinary Undergraduate Research


Corbit, Aaron


Humans have a significant impact on wildlife populations. Although not as obvious, even non-consumptive recreational activities (i.e. hiking and mountain biking) can impact wildlife. Previous research has suggested that human traffic can impact the movement patterns of the white-tailed deer (Odocoileus virginianus). Specifically, as human traffic increases, deer sightings decrease. Also, due to their crepuscular nature, deer peak activity is at dawn and dusk, yet one study reported a decrease in deer sightings in the evening that seemed to correspond to a peak in human traffic. However, the cause-and-effect relationship between human traffic and deer daily movements has yet to be fully established. In this study, we used trail cameras to examine the effect of human traffic on the abundance and daily movement patterns of O. virginianus on a private trail system on the campus of Southern Adventist University near Chattanooga, TN. Since our data generated thousands of images, we also tested the efficacy of a simple machine learning algorithm that used the Structural Similarity Index (SSIM) to help us find deer or humans within the images generated by the trail cameras. As with previous research, we found a reduction in deer observations as human traffic increased. We also found that temperature, humidity, and wind speed were inversely related to deer sightings while atmospheric pressure was directly related to deer sightings. While some aspects of our data support the hypotheses that human traffic impacts diel movement patterns of deer, other aspects are inconsistent with this hypothesis so this relationship remains unresolved. Though our machine learning methodology was not very effective, our results suggest that the use of SSIM could prove useful with further refinement. This study adds to our understanding of the ways that non-consumptive, outdoor recreational activities can impact wildlife and may help inform land use policies.