N. Gross

Temperate seminatural grasslands harbour unique biodiversity, support livestock farming through forage production, and deliver many essential ecosystem services (ESs) to human society; they are highly multifunctional. However, temperate grassland ecosystems are also among the most threatened ecosystems on earth due to land use and climate changes. Understanding how biodiversity, climate and land use intensification impact grassland multifunctionality through complex direct and indirect pathways is critical to better anticipate the future of these fragile ecosystems. Here, we evaluate how local plant species richness (SR) modulates the effect of land use intensification and climate on grassland multifunctionality (using six key ESs: biomass productivity and stability, forage quality, carbon storage, pollination and local plant rarity) in the French Massif Central, the largest grassland in Western Europe. We sampled 100 grasslands with contrasted fertilization rates and SR, over large elevational and latitudinal gradients related to variation in mean annual temperature (MAT), and drought severity (DS), two key climate change drivers predicted to increase in the future. Using a confirmatory path analysis, we found that SR was the main driver of multifunctionality. We also found significant SR × MAT and SR × fertilization interactions suggesting that warm climate and high fertilization rates alter the biodiversity–ecosystem multifunctionality relationships. Furthermore, increasing temperature and fertilization indirectly influenced multifunctionality by decreasing SR and consequent multifunctionality in warm lowland and highly fertilized grasslands compared to colder montane grasslands or less fertilized ones. DS only impacted some ES individually (e.g. forage quality). Synthesis and applications : We identified species richness (SR) as a pivotal factor mediating the effects of land use intensification and climate on multifunctionality through both direct and indirect pathways. Failing to account for changes in SR could thus bias any prediction of, or aggravate, the effects of land use intensification and climate change on ecosystem services delivery in temperate grassland ecosystems. Considering that SR, mean annual temperature and fertilization are major proxies of three main global change drivers (biodiversity loss, climate change and land use intensification) our study may help to better anticipate the effect of multiple interacting global change drivers on grassland ecosystems.