Human-wildlife coexistence in the Anthropocene requires an understanding of wildlife responses to anthropogenic change. Increases in human activity are recognized as an important stressor on wildlife, with the potential for effects distinct from, and interactive with, those from habitat loss. Two key challenges to understanding and managing the effects of human activity on wildlife have been a lack of experimental manipulations of activity and of concurrent monitoring of people and wildlife. Global changes in human activity during the Covid-19 pandemic provided quasi-experimental conditions through which to study wildlife responses, and camera traps provide a tool for standardized estimation of activity for wildlife and people. We capitalized on this opportunity by synthesizing camera trap (CT) detections of people and terrestrial mammals across more than 100 CT surveys from around the world (involving over 200 collaborating data contributors) that sampled before and during the first year of the Covid-19 pandemic. We used negative binomial generalized linear mixed models to compare activity (CT detection rates) for 293 mammal species between periods of higher vs. lower human activity, and we used a mixed-effects meta-analysis to test hypotheses about the relative influences of species traits, landscape context, and magnitude of human change on species responses.
CT detections of people highlighted the tremendous variety of changes in human activity across study areas during the pandemic, including both large increases and decreases as well as no change. Wildlife responses were equally variable, with no dominant or consistent responses across all species and sites to changes in human activity. However, trophic level emerged as the strongest predictor of variation in species responses across sites, with carnivores showing more negative responses to increasing human activity, and herbivores showing more positive responses. The degree of openness of habitat in a sampling area influenced species responses, with more negative responses to human activity in landscapes with more open habitat (i.e. greater visibility). Surprisingly, the amount of anthropogenic habitat disturbance (human footprint) had a less pronounced effect on wildlife responses. This unprecedented global “experiment” highlights that mammal responses to changing human activity are highly variable and context-specific, but that overall trends of increasing human activity in the Anthropocene have the potential to re-structure ecological communities and thus alter ecosystem processes. With continued improvements and coordination in camera trap monitoring, wildlife and land managers can more rapidly identify and mitigate negative human impacts to species and ecosystems.