This doctoral thesis explored the consequences of the abandonment of extensive grazing by mixed livestock (sheep, cattle, and horses) in the highly productive Atlantic grasslands of the Aralar Natural Park, Basque Country (northern Iberian Peninsula). All large herbivores were experimentally excluded from four field plots by using 50 × 50 m fences. A grazing plot was delineated next to each exclusion plot where herbivores grazed continuously during the growing season (May to November).
Grazing generally enhances nutrient cycling and mineralisation in productive grasslands. Soil physical properties and forage properties were measured in the field plots to understand how grazing affects soil processes. All measurements were carried out after 8–10 years of exclusion. Grazing affected nutrient cycling by modifying forage quality, soil thermal regimes, and water content. It prevented litter accumulation, which provided less insulation to the soil and enhanced mean summer temperatures and variability. Higher temperatures combined with higher forage quality and, generally, without water stress, enhanced nutrient cycling and mineralisation in grazed areas.
The effect of grazing on community physiognomy and biodiversity was then studied. Disturbance by large herbivores maintains high levels of plant diversity in productive grasslands; possibly because grazing provides an equalising mechanism that suppresses the competitive exclusion of weak plant species by dominant species. Since species competition occurs at fine spatial scales, a fine-scale, spatially explicit survey design was used. After 10 years of the cessation of herbivory, i.e. after long-term suppression of the strong equalising mechanism previously present, competitive species created large intraspecific patches and out-competed weaker species, thereby reducing species diversity.
Competitive exclusion is assumed to be strong in productive grasslands because niche stabilisation is expected to be weak (i.e. they are rich in resources). However, in one of the field plots, pH showed high heterogeneity, which exerted moderately strong stabilisation in the plant community. Using two field plots with contrasting plant niche stabilisation, the buffering effect of niche stabilisation on competitive exclusion was studied. The applied statistical techniques combined phylogenetically structured plant traits and spatially structured soil descriptors through species abundance. Under weak niche stabilisation, grasses with superior competitive ability (i.e. tall canopy with lateral vegetative spread) outcompeted dicots in all branches in the phylogenetic tree, resulting in strong reduction of species and phylogenetic diversity. However, under moderately strong niche stabilisation, competitive exclusion by superior species was counter-balanced by niche stabilisation, and resulted in a less important loss of species and phylogenetic diversity.
To conclude, given that 10 years of grazer exclusion retarded nutrient cycling and mineralisation and resulted in loss of spatial heterogeneity of floristic composition and plant diversity in the field plots, traditional grazing by mixed livestock (sheep, cattle, and horses) was a key ecological factor for maintaining soil function and plant diversity in Atlantic grasslands.