Extrapolation from research findings at the local scale to the regional and global scales pose a major challenge to simulation and projection of terrestrial biosphere in response to climate change. The temperate steppe in northern China is part of the grassland biome across the Eurasian Continent and plays an important role in regional and global carbon cycling. A manipulative experiment with experimental warming and N addition treatments was conducted along a precipitation gradient from meadow steppe in the east (MAP 455 mm), typical steppe in the middle (MAP 385 mm), and desert steppe in the west (MAP 310 mm) since April 2006. Given that the temperate steppe is water limited and global warming will increase evapotranspiration, we hypothesize that (1) the desert steppe will be more sensitive to temperature increase than the meadow and typical steppe, (2) global warming may pose positive impacts on the meadow steppe, but negative responses are expected in the desert steppe.

Results/Conclusions

Plant species richness was significantly reduced by 10.1% under warming in the desert steppe whereas no response was detected in the meadow steppe or the typical steppe over the 4 years from 2006 to 2009. Experimental warming significantly decreased above- and belowground net primary productivity (ANPP and BNPP) by 8.3% and 15.4%, respectively, in the desert steppe in the first 3 years (2006-2008). In contrast, BNPP showed positive responses (13.7%) to warming in the meadow steppe. The observations support our hypotheses. N addition did not affect plant diversity, but significantly stimulated both ANPP (6.3-25.3%) and BNPP (18.2-46.5%). Interactions of warming and N addition were observed to affect plant species richness only in the typical steppe. Site-dependence of plant diversity and ecosystem productivity to experimental warming observed in the temperate steppe are critical for model simulation and projection of climate-terrestrial biosphere feedback at the regional and global scale.