The moon’s influence on the activity of tropical forest mammals

Changes in lunar illumination alter the balance of risks and opportunities for animals, influencing activity patterns and species interactions. We examine if and how terrestrial mammals respond to the lunar cycle in some of the darkest places: the floors of tropical forests. We analyzed long-term camera trapping data on 86 mammal species from 17 protected forests on three continents. Conservative categorization of activity during the night revealed pronounced avoidance of moonlight (lunar phobia) in 12 species, compared with pronounced attraction to moonlight (lunar philia) in only 3 species. However, half of all species in our study responded to lunar phases, either changing how nocturnal they were, altering their overall level of activity, or both. Avoidance of full moon was more common, exhibited by 30% of all species compared with 20% of species that exhibited attraction. Nocturnal species, especially rodents, were over-represented among species that avoided the full moon. Artiodactyla was more prominent among species attracted to the full moon. Our findings indicate that lunar phases influence animal behavior even beneath the forest canopy. Such impacts may be exacerbated in degraded and fragmented forests. Our study offers a baseline representing relatively intact and well-protected contexts together with an intuitive approach for detecting activity shifts in response to environmental change.

Methods
We derived observations of mammal activity in protected tropical forests from camera trap data collected as part of the Tropical Ecology Assessment and Monitoring (TEAM) Network (1). Following a common protocol (2), cameras were deployed between 2008 and 2017 throughout 17 protected areas in Indomalaya, the Neotropics, and the Afrotropics. The number of years of deployment varied between protected areas (2 to 10 years; mean = 6.8 years), as did the number of locations sampled (60-90 camera trap locations; total: 1062). Spatial configuration and deployment were standardized, with cameras configured in either a 1x1km or 2x2km regular grid, at a height of approximately 30-50 centimeters off the ground. On average, cameras were active for 33.2 days (SD=7.5). However, cameras were rotated sequentially until all sites were sampled within the wider sampling season. As a result, multiple lunar cycles are recorded at each protected area within a sampling season. For additional information about camera trapping protocols and species identification, see Rovero and Ahumada (1). In this analysis, we included more than 2.1M photographs of 86 mammal species, i.e., species with ≥ 25 detection events (number of 15-minute intervals with at least one detection at a camera trap) during the night (total across all protected areas; Electronic Supplementary Materials Tables S1-S3). Observations associated with species identifications that were flagged as uncertain were excluded from the analysis. Due to concerns about identification, species in the genus Tragulus were considered jointly (Tragulus sp.). Observations were aggregated and prepared as input for Bayesian models as described in the README file.

References:

1. Rovero F, Ahumada J. The Tropical Ecology, Assessment and Monitoring (TEAM) Network: An early warning system for tropical rain forests. Sci Total Environ [Internet]. 2017;574:914–23. Available from: https://www.sciencedirect.com/science/article/pii/S0048969716320769

2. Jansen PA, Ahumada J, Fegraus E, O’Brien T. TEAM: a standardised camera trap survey to monitor terrestrial vertebrate communities in tropical forests. Camera Trapp Wildl Res Manag. 2014;263–70.