Carbon-13 input and turn-over in a pasture soil exposed to long-term elevated atmospheric CO 2 CHRISVANKESSEL,*JENNIFERNITSCHELM,²WILLIAMR.HORWATH,³ DAVIDHARRIS,§FRANCESWALLEY,²ANDREASLU È SCHER¶and UELIHARTWIG¶ *Department of Agronomy and Range Science, University of California-Davis, CA 95616, USA, ²Department of Soil Science, University of Saskatchewan, Saskatoon, Canada S7N 5A8; ³Land, Air and Water Resources, University of California-Davis, Davis, CA 95616, USA; §Stable Isotope Facility, University of California-Davis, Davis, CA 95616, USA; ¶Institute of Plant Sciences, Swiss Federal Institute of Technology, Zu Èrich, Switzerland Abstract The impact of elevated CO 2 and N-fertilization on soil C-cycling in Lolium perenne and Trifolium repens pastures were investigated under Free Air Carbon dioxide Enrichment (FACE) conditions. For six years, swards were exposed to ambient or elevated CO 2 (35 and 60Pa pCO 2 ) and received a low and high rate of N fertilizer. The CO 2 added in the FACE plots was depleted in 13 C compared to ambient (D ±40½) thus the C inputs could be quanti®ed. On average, 57% of the C associated with the sand fraction of the soil was `new' C. Smaller proportions of the C associated with the silt (18%) and clay fractions (14%) were derived from FACE. Only a small fraction of the total C pool below 10cm depth was sequestered during the FACE experiment. The annual net input of C in the FACE soil (0±10cm) was estimated at 4.6 6 2.2 and 6.3 6 3.6 (95% con®dence interval) Mgha ±1 for T. repens and L. perenne, respectively. The maximum amount of labile C in the T. repens sward was estimated at 8.3 6 1.6Mgha ±1 and 7.1 6 1.0Mgha ±1 in the L. perenne sward. Mean residence time (MRT) for newly sequestered soil C was estimated at 1.8years in the T. repens plots and 1.1years for L. perenne. An average of 18% of total soil C in the 0±10cm depth in the T. repens sward and 24% in the L. perenne sward was derived from FACE after 6years exposure. The majority of the change in soil d 13 C occurred in the ®rst three years of the experiment. No treatment effects on total soil C were detected. The fraction of FACE-derived C in the L. perenne sward was larger than in the T. repens sward. This suggests a priming effect in the L. perenne sward which led to increased losses of the old C. Although the rate of C cycling was affected by species and elevated CO 2 , the soil in this intensively managed grassland ecosystem did not become a sink for additional new C. Keywords: annual C input, carbon-13, elevated CO 2 , FACE, labile soil organic matter pool, long-term exposure, particle size, pasture Received 15 February 1999; resubmitted and accepted 11 May 1999 Introduction Combustion of fossil fuels released an estimated 5Pgy ±1 in 1990 of C to the atmosphere, primarily as CO 2 , and emissions continue to increase (Taylor & Lloyd 1992). Deforestation driven by increasing populations and land degradation (Cole etal. 1993), results in large C ¯uxes to the atmosphere of approximately 2Pg Cy ±1 (Post etal. 1990). These additional inputs are expected to increase atmospheric CO 2 concentration to twice its preindustrial value of 280ppm by 2050 (Keeling 1993). Increased atmospheric CO 2 can lead to greater plant biomass production, and some researchers have sug- gested that suf®cient soil organic matter will be formed *Correspondence: tel. +1/530-752-4377, e-mail cvankessel@ucdavis.edu Global Change Biology (2000) 6, 123±135 # 2000 Blackwell Science Ltd. 123