Multi-stress-affected sandy grasslands: Livestock-grazing as a tool for nature conservation under the impact of drought events and rabbit population fluctuations?

Species-rich inland sandy grasslands in central Europe are of high nature conservation value and highly endangered. Livestock grazing is thought to be beneficial for maintaining their value, but data on its long-term effects are scarce. Intense stress factors (drought events, rabbit overpopulation) occur episodically and may interact with the grazing effects in a complex manner. We carried out a long-term study to disentangle these factors and their possible impacts on the vegetation. The study site was situated in the Upper Rhine valley in Germany. The dominant plant communities were the threatened vegetation types Koelerion glaucae and Armerio-Festucetum trachyphyllae. After some years of the site lying fallow, grazing management by sheep was established in 2000 and in 2002 also by donkeys. For 17 years (2000-2016), we studied the vegetation using a stratified randomised block design of 14 fenced plots with livestock exclusion and 14 adjacent livestock-grazed permanent plots (of 25 m(2) each). Due to the increasing rabbit population, in 2005, we established additional plots (n = 6, with the exclusion of livestock + rabbits). Releves were sampled yearly and analysed using ordination, linear regressions and mixed linear models. The following results were obtained: 1. Drought events led to a marked drop in species richness. Most of the species recovered within the next year owing to the soil seed/bud bank. However, some species, particularly perennial graminoids, were impaired more permanently. 2. Rabbit overpopulation affected structural variables: the open soil area increased and litter cover decreased. After the decline of the rabbit population, these variables showed reversed trends again. 3. The system studied was susceptible to stress factors, and, therefore, many variables were characterised by inter-annual fluctuations. However, the plant community types also showed pronounced resilience and general stability if grazers were present. 4. Plots with rabbit + livestock exclusion experienced the encroachment of graminoids, litter accumulation and the loss of species (7 species on average). 5. In this stress -driven system, grazing by livestock had clear effects in the last 4 years. Initially, the grazing intensity was too low; in the middle phase, rabbit-grazing was superimposed on the livestock effects. Afterwards, the livestock-grazed plots showed higher proportions of small species, higher target species ratios and, in the Armerio-Festucetum stands, also more Red-list species. Unlike the strongly fluctuating population sizes of rabbits, livestock impact can be regulated and is, thus, a necessary tool for nature conservation. The study demonstrates the necessity for long-term studies to disentangle the complex interactions of stress factors and ecosystem responses.