Herbivores are considered to have significant impact on their environment. This impact is usually attributed to trophic relationships between the herbivores and plants. However, herbivores affect ecological systems not only by exploitation of plant biomass, but also by functioning as ecosystem engineers. We empirically studied the impact of goats browsing on three levels of organization. We assessed the effects of goat grazing on landscape structure and how the modulation of landscape patchiness affects processes at ecosystem and community levels. We addressed this question in two-phase mosaic, composed of open and woody patches, and investigated how grazing impacts differ between different patch types. Our model system was the Mediterranean woodland in the Carmel Mountain, Israel. We used five plots of 1000 m² with goat grazing and five plots without grazing. Plant species richness and composition were sampled in each plot and in two adjacent patch-types: woody and open (herbaceous). We collected data on environmental conditions for each patch type in each plot, with and without grazing. Processing of Helium balloon photos was used for spatial analysis of landscape structure by FRAGSTAT software.

Results/Conclusions

We found that grazing modified the structure of woody patches and the whole landscape by soil trampling, braking branches and woody vegetation consumption. Modification of landscape patchiness affected ecosystem processes by changing material flow and affecting availability of resources in both woody and open patch types. Grazing impact on plant community was patch type dependent: in open patches grazing reduced plant species richness, as a result of exploitation of herbaceous vegetation. In woody patches, grazing increased plant species richness, as a result of improving abiotic conditions by engineering. At a higher spatial scale (plot), grazing did not change plant species richness because the negative effect of consumption in the open patches was balanced by the positive effect of engineering in the woody patches. Since the impact of herbivores as exploiters or engineers is determined by the character of the patches, the number, size and distribution of open and woody patch-types in the landscape determine the relative effect of grazing as exploiters or modulators. We conclude that analyzing grazing effects on plant community should take into account the structure of the two-phase landscape, because of the different effects of grazing in different patch types. Indirect impact through modifying processes at landscape and ecosystem levels is an important factor by which grazing affects plant community.