Human activities can influence behaviors of predators and prey, as well as predator-prey interactions. Using camera trap data, we investigated whether or to what extent human activities influenced behaviors of predators (tigers and leopards) and prey (sambar deer, spotted deer, wild boar, barking deer), and predator-prey interactions in the Barandabhar Corridor Forest (BCF), Chitwan District, Nepal. A multispecies occupancy model revealed that the presence of humans altered the conditional occupancy of both prey and predator species. Specifically, the conditional occupancy probability of prey was substantially higher (ψ = 0.91, CI = 0.89–0.92) when humans were present than when humans were absent (ψ = 0.68, CI = 0.54–0.79). The diel activity pattern of most prey species overlapped strongly with humans, whereas predators were generally more active when humans were absent. Finally, the spatio-temporal overlap analysis revealed that human-prey interactions (i.e., the probability that both humans and prey species being present on the same grid at the same hourly period) was ~3 times higher (10.5%, CI = 10.4%–10.6%) compared to spatio-temporal overlap between humans-predators (3.1%, CI = 3.0%–3.2%). Our findings are consistent with the human-shield hypothesis and suggest that ungulate prey species may reduce predation risk by using areas with high human activities.

Camera trap surveys were conducted between December 2015 and January 2016. The study area (Barandabhar Corridor Forest, BCF, area ~88 sq km) was divided into 88 grids measuring 1 X 1 km each. Unbaited camera traps were strategically placed at each grid, along forest roads and trails, based on signs of prey and predator presence. Multiple camera trap models were used. The traps were set up 0.5 m above ground, 4–5 m apart on both sides of the trails, and remained active for a minimum of 20 days, totaling 1,760 trap nights.

The dataset includes four prey species (sambar deer, spotted deer, wild boar, and barking deer), two predator species (tigers and leopards), and humans. Multispecies occupancy models were used to analyze the detection history matrix, considering prey and predator guilds. The model compared the probability of occurrence for different species combinations, both in the presence and absence of humans.

Temporal overlap analysis examined diel activity patterns between humans and prey, as well as humans and predators. Coefficient of overlap (Δ) was calculated to determine the degree of overlap between two species' activity patterns. The overlap was categorized as low, moderate, or high based on the Δ value. Bootstrap resampling was used to obtain confidence intervals for each overlap index.

Spatio-temporal overlap analysis was performed to assess the overlap in space and time between species. A matrix of species detection per hour was created for each grid, indicating the presence of predators or prey during specific hourly intervals. Proportions of grids were calculated where species were detected alone, where two species overlapped, and where all three species were active. Bootstrap resampling was used to estimate standard errors and confidence intervals for the spatio-temporal overlap.

Microsoft Excel, and RStudio to run the R-code given in the '.txt' format.