Habitat impact assessment detects spatially driven patterns of grazing impacts in habitat mosaics but overestimates damage

Many habitats of conservation importance are grazed by large herbivores, but spatial variation in grazing intensity can complicate management planning. We tested the effect of local herbivore density and the proximity of alternative preferred plant communities on the distribution of large herbivore impacts on vegetation. We analysed Habitat Impact Assessment (HIA) data collected from four different habitats of conservation value (Species-rich grassland; Blanket Bog; Dry Heath; Wet Heath) on the Isle of Rum National Nature Reserve. Species-rich grassland (SRG) is the preferred plant habitat for the main herbivores present (red deer, cattle and ponies). Grazing impacts on the three less preferred habitats were positively correlated with the proportion of the preferred habitat within 1 km, explaining 11-16% of the variance. There was an unexpected negative correlation between the proportion of preferred habitat within 1 km and the grazing impacts on SRG, although the variation explained was modest (5.6%). The distribution of grazing on SRG was strongly negatively correlated with elevation. Local herbivore density was weakly, positively correlated with grazing impact on Dry Heath. Herbivore density at the level of management unit was either not associated or only weakly associated with impacts; the identity of the management unit generally explained more variance in impacts. The results suggest that the HIA method of data collection was suitable to assess the potential for conflicts in management goals for adjacent habitats with different herbivore grazing preferences with Dry Heath most strongly affected by this spill-over effect. For three habitats that could be compared, HIA data suggested higher grazing impacts than a fixed quadrat method focussed on physiologically relevant measures and consequently the HIA approach may be too sensitive in assessing habitat condition.