The effects of herbivores on ecosystem functioning belowground are not as well understood as aboveground. To find global patterns in soil biological responses to vertebrate herbivores, we performed a meta-analysis of field exclosure experiments in which soil animal or microbial responses to mammalian herbivores had been measured. We tested whether the responses may be predicted based on biotic and environmental controls such as ecosystem type (e.g. grassland, forest, etc.), climate, and herbivore identity. Furthermore, we assessed whether our quantitative findings fit two frameworks of soil biological responses to aboveground herbivores: Bardgett & Wardle (2003), who generalized soil decomposer responses based on site fertility and level of herbivory; and Schrama et al. (2013), who generalized soil nitrogen mineralization responses based on soil texture and moisture and on herbivore physical disturbance. Finally, we sought to identify the most critical research gaps.
Soil biological responses to mammalian herbivores varied among biomes. Climate and herbivore identity were identified as stronger controls than ecosystem type, and climate categories performed better than simpler metrics such as precipitation and temperature. Responses varied across and within trophic levels in the soil food web, suggesting shifts in soil functioning when herbivores are gained or lost aboveground. While responses to herbivores in low-productive ecosystems were often negative as predicted by Bardgett & Wardle, also the responses in productive ecosystems were rarely positive. The outline by Schrama and co-authors could not be rigorously tested due to scarcity of soil physical data; however, our finding of soil biological responses becoming negative with increasing herbivore body size is consistent with the role of herbivore physical disturbance in that framework. To complement previous theory and link it with our findings, we propose a new conceptual diagram, in which soil biological responses switch from weakly positive to strongly negative with increasing herbivore body size and/or increasing climate severity. Mechanisms behind the observed responses and critical research gaps that still hinder predictions at the global level are discussed.