Using the dynamics of productivity and precipitation-use efficiency to detect state transitions in Eurasian grasslands

In the face of accelerated global dryland expansion and grassland degradation, signaling grassland ecosystem state transitions is an ongoing challenge in ecology. However, there is still a lack of effective indicators and understanding of the mechanisms of grassland ecosystem state transitions at the continental scale. Here, we propose a framework that links ecosystem function-based indicators and critical slowing down (CSD) theory to reveal grassland state transitions. Across precipitation gradients, we quantified the statistical characteristics and spatial patterns in ANPP and PUE dynamics (variability, asymmetry, and sensitivity to precipitation and temperature) in Eurasian grasslands. We show that the CVANPP, CVPUE, A(ANPP), A(PUE), SPUE-P, and SANPP-P of temperate steppes were significantly higher than those of alpine steppes, while the SPUE-T and SANPP-T were the inverse. In temperate grasslands, A(ANPP), A(PUE), and SANPP-P indicated the transition of typical steppes, and CVANPP, A(PUE), and SPUE-T indicated the transition from meadow to typical steppes. In alpine grasslands, A(PUE) indicated the transition between alpine deserts and alpine steppes, and A(ANPP) and SANPP-P indicated the transition between alpine steppes and meadow steppes. The interannual variability of precipitation strongly affected xerophyte proportion and demographic processes, which control state transitions in low-resilience grasslands. Community structures and limiting factors (nutrient, light, and/or temperature) regulate state transitions in high-resilience grasslands. Our results demonstrate that function-based indicators are predictive of impending state transitions of temperate and alpine grasslands, highlighting the complementation of ANPP and PUE dynamics that have the potential for predicting grassland ecosystem regime shifts and their underlying mechanisms.