Using remote cameras to measure seasonal molts

Obtaining accurate data on seasonal timing of lifecycle events in the wild is critical for many aspects of ecological research. However, characterizing such phenological processes is difficult, expensive, and time consuming. Remote camera traps are increasingly used in ecology, yet their potential to study key phenological traits in animal populations has been largely unexplored. Here, we examine the potential of remote camera traps to measure the progression of seasonal molts in mammals. We evaluated the accuracy of trained observers to classify the stage of molt from camera-trap images and identified factors that increase the accuracy of this method in a common, color molting mammal, the snowshoe hare (Lepus americanus). Our results showed that images taken by remote camera traps can be used to classify the stage of color molt with relatively high accuracy (i.e., 84%). Observers achieved the highest accuracy when using a classification protocol with fewer molt categories, and from images acquired during the day. We also found that hare body position in the image, and whether the hare was moving or still had small influences on observer classification accuracy. Camera model had negligible effect on accuracy. Overall, our results suggest that camera traps can be used to classify molt progression to measure molt phenology in the wild. Because many camera-trap studies are ongoing around the world, images of species that undergo distinguishable seasonal molts could be pooled across studies to characterize molt phenology on local and global scales. In much the same way that remote cameras have revolutionized the study of distribution, abundance, and behavior of some animal populations, so too can remote camera images transform our understanding of key phenological processes across space, time, and taxa.