The fluxes of CO2 from grazed and fenced temperate steppe during two drought years on the Inner Mongolia Plateau, China

The CO2 flux was measured by the eddy covariance method on a temperate Leymus chinensis steppe over a period of 17 months spanning two consecutive growing seasons. The amount of precipitation was nearly normal, but it was low in the early and high in the late growing period in 2006. In the 2007 growing season, the amount of precipitation was about 45% less than the multi-year average and more evenly distributed. Comparisons were made between a moderately grazed site and a 28-year-old fenced site. The maximum instantaneous CO2 release and uptake rates were 0.12 (May) and 0.11 mg CO2 m(-2) s(-1) (July) at the fenced site, and 0.11 and 0.16 mg CO2 m(-2) s(-1) (both in July) at the grazed site. In both growing seasons, the grazed site always had a higher daily uptake rate or lower release rate than the fenced site. The grazed site was a CO2 sink during the growing season of 2007 and a CO2 source in the growing season of 2006, whereas the fenced site was a CO2 source in both seasons. Lower precipitation decreased CO2 loss during the growing season more in the grazed site than in the fenced site, mainly because of depression of total ecosystem respiration (R-e) in the former and stimulation in the latter. During the dormant season (from October to April), the fenced and grazed sites released 60.0 and 32.4 g of C per m(2), respectively. Path analysis showed that temperature had the greatest effect on daily variation of ecosystem CO2 exchange during the growing seasons at the two study sites. The results suggest that decrease of precipitation and/or increase of temperature will likely promote C loss from L chinensis steppes, whether fenced or grazed, and that a grazed site is more sensitive. (C) 2011 Elsevier B.V. All rights reserved.