Concomitant changes of annual precipitation and its seasonal distribution within the context of global climate change have dramatic impacts on aboveground net primary productivity (ANPP) of grassland ecosystems. In this study, combining remote sensing products with in situ measurements of ANPP, we quantified the effects of mean annual precipitation (MAP) and precipitation seasonal distribution (PSD) on the spatial variations in ANPP along a climate gradient in Eurasian temperate grassland. Our results indicated that ANPP increased exponentially with MAP for the entire temperate grassland, but linearly for a specific grassland type, i.e. the desert steppe, typical steppe, and meadow steppe from arid to humid regions. The slope of the linear relationship appeared to be steeper in the more humid meadow steppe than that in the drier typical and desert steppes. PSD also had significant effect on the spatial variations in ANPP. It explained 39.4% of the spatial ANPP for the entire grassland investigated, being comparable with the explanatory power of MAP (40.0%). On the other hand, the relative contribution of PSD and MAP is grassland type specific. MAP exhibited a much stronger explanatory power than PSD for the desert steppe and the meadow steppe at the dry and wet end, respectively. However, PSD was the dominant factor affecting the spatial variation in ANPP for the median typical steppe. Our results imply that altered pattern of PSD due to climate change may be as important as the total amount in terms of effects on ANPP in Eurasian temperate grassland.

Grazing exclusion (GE) is considered to be an effective approach to restore degraded grasslands and to improve their carbon (C) sequestration. However, the C dynamics and related controlling factors in grasslands with GE have not been well characterized. This synthesis examines the dynamics of soil C content and vegetation biomass with the recovery age through synthesizing results of 51 sites in grasslands in China. The results illustrate increases in soil C content and vegetation biomass with GE at most sites. Generally, both soil C content and vegetation biomass arrive at steady state after 15 years of GE. In comparison, the rates of increase in above-and belowground biomass declined exponentially with the age of GE, whereas soil C content declined in a milder (linear) way, implying a lagged response of soil C to the inputs from plant biomass. Mean annual precipitation (MAP) and the rate of soil nitrogen (N) change were the main factors affecting the rate of soil C content change. MAP played a major role at the early stage, whereas the rate of soil N change was the major contributor at the middle and late stages. Our results imply that the national grassland restoration projects in China may be more beneficial for C sequestration in humid regions with high MAP. In addition, increased soil N supply to grasslands with GE at the latter recovery stage may enhance ecosystem C sequestration capacity.