Biodiversity can stabilise productivity through different mechanisms, such as asynchronous species responses to environmental variability and species stability. Global changes, like intensified drought, could negatively affect species richness, species asynchrony and species stability, but it is unclear how changes in these mechanisms will affect the stability of above‐ground primary productivity (ANPP) across ecosystems. We studied the effects of a 4‐year extreme drought on ANPP stability and the underlying mechanisms (species richness, species asynchrony and species stability) across six grasslands in Northern China. We also assessed the relative importance of these mechanisms in determining ANPP stability under extreme drought. We found that extreme drought decreased ANPP stability, species richness, species asynchrony and species stability across the six grasslands. However, structural equation modelling revealed that species asynchrony, not species richness or species stability, was the most important mechanism promoting stability of ANPP, regardless of drought across the six grasslands. Synthesis. Our results suggest that species asynchrony, not species richness and species stability, consistently buffers ecosystem stability against extreme drought across and within grasslands spanning a broad precipitation gradient. Thus, species asynchrony may be a more general mechanism for promoting stability of ANPP in grasslands in the face of intensified drought.

Nutrient enrichment can simultaneously increase and destabilise plant biomass production, with co‐limitation by multiple nutrients potentially intensifying these effects. Here, we test how factorial additions of nitrogen (N), phosphorus (P) and potassium with essential nutrients (K+) affect the stability (mean/standard deviation) of aboveground biomass in 34 grasslands over 7 years. Destabilisation with fertilisation was prevalent but was driven by single nutrients, not synergistic nutrient interactions. On average, N‐based treatments increased mean biomass production by 21–51% but increased its standard deviation by 40–68% and so consistently reduced stability. Adding P increased interannual variability and reduced stability without altering mean biomass, while K+ had no general effects. Declines in stability were largest in the most nutrient‐limited grasslands, or where nutrients reduced species richness or intensified species synchrony. We show that nutrients can differentially impact the stability of biomass production, with N and P in particular disproportionately increasing its interannual variability.

Recurrent droughts are an inevitable consequence of climate change, yet how grasslands respond to such events is unclear. We conducted a 6-year rainfall manipulation experiment in a semiarid grassland that consisted of an initial 2-year drought (2015-2016), followed by a recovery period (2017)(2018) and, finally, a second 2-year drought (2019)(2020). In each year, we estimated aboveground net primary productivity (ANPP), species richness, community-weighted mean (CWM) plant traits, and several indices of functional diversity. The initial drought led to reduced ANPP, which was primarily driven by limited growth of forbs in the first year and grasses in the second year. Total ANPP completely recovered as the rapid recovery of grass productivity compensated for the slow recovery of forb productivity. The subsequent drought led to a greater reduction in total ANPP than the initial drought due to the greater decline of both grass and forb productivity. The structural equation models revealed that soil moisture influenced ANPP responses directly during the initial drought, and indirectly during the subsequent drought by lowering functional diversity, which resulted in reduced total ANPP. Additionally, ANPP was positively influenced by CWM plant height and leaf nitrogen during the recovery period and recurrent drought, respectively.Overall, the greater impact of the second drought on ecosystem function than

No abstract

1. Grasslands are expected to experience droughts of unprecedented frequency and magnitude in the future. Characterizing grassland responses and recovery from drought is therefore critical to predict the vulnerability of grassland ecosystems to climate change. Most previous studies have focused on ecosystem responses during drought while investigations of post-drought recovery are rare.Few studies have used functional traits, and in particular bud or clonal traits, to explore the mechanisms underlying grassland responses to and recovery from drought.2. To address this issue, we experimentally imposed a four-year drought in a C 3dominated grassland in northeastern China and monitored recovery for 3 years post-drought. We investigated the immediate and legacy effects of drought on total above-ground net primary productivity (ANPP), ANPP of functional groups (rhizomatous grasses, bunch grasses and forbs), and how the legacy effects were driven by plant species diversity, clonal traits and vegetative traits.3. We found that drought progressively reduced total ANPP over the 4-year period.The reductions in total ANPP in the first and third drought years were caused by the decrease in ANPP of bunch grasses only, while that of the second year was caused by declines in ANPP of bunch grasses and forbs, and the fourth year decline was linked to all three functional groups. The post-drought recovery of ANPP, which occurred despite the continued loss of plant species diversity, was mainly driven by rapid recovery of rhizomatous and bunch grasses, which compensated for the slow response by forbs. The rapid post-drought recovery of