Summer grazing by three livestock species at moderate intensity enhances primary productivity and drives community divergence in a semi‐arid steppe

Abstract

AbstractAimsThe general effects of grazing by large herbivores on plant species composition and productivity are well known. Meanwhile, how different livestock species affect community structure is not fully understood. In northern China, livestock species composition is changing rapidly and hence, it is imperative to understand the resulting effects on grassland productivity and composition.LocationXilingol, Inner Mongolia, China.MethodsWe studied how grazing by three major livestock species affects typical steppe dominated by Leymus chinensis, Stipa grandis, Cleistogenes squarrosa and Carex korshinskyi, in a 4‐year field experiment using a random block‐design with four treatments (sheep, goat, cattle, no grazing), under summer grazing at moderate intensity. We calculated above‐ground net primary productivity (ANPP) and a selectivity index (SI) based on biomass to quantify livestock forage preferences.ResultWe found that (i) sheep preferred Leymus chinensis and mixed subordinate species, cattle preferred Stipa grandis, and goat preferred Carex korshinskyi and Cleistogenes squarrosa. (ii) Consistent with these preferences, sheep increased the relative biomass of Stipa grandis, but reduced that of Carex korshinskyi; cattle reduced the relative biomass of Stipa grandis and Carex korshinskyi but increased that of mixed subordinate species; and goats reduced the relative biomass of Cleistogenes squarrosa and Carex korshinskyi but increased that of mixed subordinate species. (iii) Overall, grazing significantly increased ANPP, mainly due to an increase in ANPP of Leymus chinensis and subordinate species, at the cost of Stipa grandis.ConclusionsSummer grazing by different livestock species at moderate intensity can increase ANPP, while livestock forage preferences can drive divergent changes in species composition. Our results also imply that livestock composition can be used as a mean to manage vegetation dynamics.