Long-term heavy grazing increases community-level foliar fungal diseases by shifting plant composition.
In grasslands, the interactions of foliar fungal diseases and their host plants are largely dependent on grazing by large herbivores. However, the relative importance of direct (i.e. pathogen removal) and indirect effects (i.e. via changes in plant community composition) of long-term grazing on foliar fungal diseases remains largely unexplored, especially under varied grazing intensities. We conducted a 13-year experiment to explore the effects of grazing intensity on foliar fungal diseases at both plant population and community levels in a semi-arid grassland. We quantified the contributions of direct and indirect effects of long-term grazing on community pathogen load. At the population level, the severity of five rusts and five powdery mildews decreased significantly as grazing intensity increased, whereas two leaf spots increased significantly in severity with increasing grazing intensity. Similarly, at the community level, the pathogen load of rusts and powdery mildews was negatively related to the increase in grazing intensity, whereas the pathogen load of leaf spots was positively related to grazing intensity. Overall, heavy grazing (i.e. 8.7 sheep/ha) significantly increased community pathogen load. Our SEM analysis showed that grazing indirectly increased the pathogen load of leaf spots by increasing disease proneness. Grazing decreased the pathogen load of rusts, but this could not be explained by changes in disease proneness. Overall, the indirect effects via changes in community composition of hosts resulting in the increase in community disease proneness outweighed the direct effect of grazing on community pathogen load. Synthesis and applications. Our study provides the first evidence that long-term heavy grazing can indirectly increase community pathogen load by increasing the abundances of grazing-tolerant hosts and decreasing the abundances of grazing-intolerant hosts. These results provide empirical evidence that the pathogen load of foliar fungal diseases in grasslands can depend on the community context of hosts, which can, in turn, be controlled by large herbivores. We recommend that infectious diseases are considered when predicting the responses of grassland ecosystems to anthropogenic activities. Maintaining light to moderate grazing intensity or establishing an appropriate non-grazing period could be an effective way to control foliar fungal diseases in grasslands.