Skip to main content

Data from: On the scaling of activity in tropical forest mammals

Cite this dataset

Cid, Bruno et al. (2020). Data from: On the scaling of activity in tropical forest mammals [Dataset]. Dryad.


Activity range – the amount of time spent active per day – is a fundamental aspect contributing to the optimization process by which animals achieve energetic balance. Based on their size and the nature of their diet, theoretical expectations are that larger carnivores need more time active to fulfil their energetic needs than do smaller ones and also more time active than similar-sized non-carnivores. Despite the relationship between daily activity, individual range and energy acquisition, large-scale relationships between activity range and body mass among wild mammals have never been properly addressed. This study aimed to understand the scaling of activity range with body mass, while controlling for phylogeny and diet. We built simple empirical predictions for the scaling of activity range with body mass for mammals of different trophic guilds and used a phylogenetically controlled mixed model to test these predictions using activity records of 249 mammal populations (128 species) in 19 tropical forests (in 15 countries) obtained using camera traps. Our scaling model predicted a steeper scaling of activity range in carnivores (0.21) with higher levels of activity (higher intercept), and near-zero scaling in herbivores (0.04). Empirical data showed that activity ranges scaled positively with body mass for carnivores (0.061), which also had higher intercept value, but not for herbivores, omnivores and insectivores, in general, corresponding with the predictions. Despite the many factors that shape animal activity at local scales, we found a general pattern showing that large carnivores need more time active in a day to meet their energetic demands.


Camera-trap dataset

We obtained empirical data on activity range from cameratrapping surveys in 19 tropical forests, distributed over four continents. Seventeen sites were part of the Tropical Ecology Assessment and Monitoring Network, and two additional sites were located in South America. All sites were surveyed by 50 nonbaited cameras, spaced 2 km from each other. The camera grid was set for 30 consecutive days, during dry seasons, totalizing 1500 camera-days, according to the standardized TEAM protocol. We trimmed the photographic captures (henceforth ‘records’) in each database to meet the minimum of 1-h interval between consecutive records of the same species at the same camera site to reduce temporal dependence. At each site, we only used species that had a minimum of 30 independent records, and excluded Homo sapiens and domesticated animals (as all surveyed sites were inside protected areas, the influence of people and domestic animals on the wildlife activity can be considered negligible). We defined the minimum of 30 records per species after performing analysis of sampling sufficiency through bootstrapping in order to estimate activity ranges.