No evidence for the consistent effect of supplementary feeding on home range size in terrestrial mammals

Food availability and distribution are key drivers of animal space use. Supplemental food provided by humans can be more abundant and predictable than natural resources. It is thus believed that supplementary feeding modifies the spatial behaviour of wildlife. Yet, such effects have not been tested quantitatively across species. Here, we analysed changes in home range size due to supplementary feeding in 23 species of terrestrial mammals using a meta-analysis of 28 studies. Additionally, we investigated the moderating effect of factors related to i) species biology (sex, body mass, taxonomic group), ii) feeding regimen (duration, amount, purpose), and iii) methods of data collection and analysis (source of data, estimator, spatial confinement). We found no consistent effect of supplementary feeding on changes in home range size. While an overall tendency of reduced home range was observed, moderators varied in the direction and strength of the trends. Our results suggest that multiple drivers and complex mechanisms of home range behaviour can make it insensitive to manipulation with supplementary feeding. The small number of available studies stands in contrast with the ubiquity and magnitude of supplementary feeding worldwide, highlighting a knowledge gap in our understanding of the effects of supplementary feeding on ranging behaviour.

Methods
We used Web of Science (WoS) and Scopus, two publisher-independent global citation databases, to identify published articles that analysed the effect of supplementary feeding on home range size in terrestrial mammals.We extracted home range size, its standard deviation (SD), and sample size from each publication for the experimental group of animals with supplementary feeding and the control group without.For each home range comparison, we compiled information about the species, its taxonomic group (rodent, carnivore, and ungulate), individual's sex, supplementary food amount (limited or ad libitum) and feeding duration, spatial confinement (free-ranging or enclosure), source of spatial data (telemetry or capture-mark-recapture), and home range estimator (kernel density estimation (KDE) or minimum convex polygon (MCP)). Moreover, we divided the publications into intentional and unintentional feeding studies.To measure the change of the home range size from no-feeding to feeding treatment, we used the Hedges'g estimator of the effect size.Meta-analytic mixed-effects models, fitted with the function rma.rm of the metafor package, were used to examine the effects of supplementary feeding on home range size.