Long-term steady state ¹³C labelling to investigate soil carbon turnover in grasslands

We have set up a facility allowing steady state 13CO2 labeling of short stature vegetation (12 m2) for several years. 13C labelling is obtained by scrubbing the CO2 from outdoors air with a self-regenerating molecular sieve and by replacing it with 13C depleted (-34.7±0.03‰) fossil-fuel derived CO2 The facility, which comprises 16 replicate mesocosms, allows to trace the fate of photosynthetic carbon in plant-soil systems in natural light and at outdoors temperature. This method was applied to the study of soil organic carbon turnover in temperate grasslands. We tested the hypothesis that a low disturbance by grazing and cutting of the grassland increases the mean residence time of carbon in coarse (>0.2 mm) soil organic fractions. Grassland monoliths (0.5×0.5×0.4 m) were sampled from high and low disturbance treatments in a long-term (14 yrs) grazing experiment and were placed during two years in the mesocosms. During daytime, the canopy enclosure in each mesocosm was supplied in an open flow with air at mean CO2 concentration of 425 µmol mol-1 and d13C of -21.5±0.27‰. Fully labelled mature grass leaves reached a d13C of -40.8 (±0.93) and -42.2‰ (±0.60) in the low and high disturbance treatments, respectively, indicating a mean 13C labelling intensity of 12.7‰ compared to unlabelled control grass leaves. After two years, the delta 13C value of total soil organic matter above 0.2 mm was reduced in average by 7.8‰ in the labelled monoliths compared to controls. The isotope mass balance technique was used to calculate for the top (0–10 cm) soil the fraction of 13C labelled carbon in the soil organic matter above 0.2 mm (i.e. roots, rhizomes and particulate organic matter). A first order exponential decay model fitted to the unlabelled C in this fraction shows an increase in mean residence time from 22 to 31 months at low compared to high disturbance. A slower decay of roots, rhizomes and particulate organic matter above 0.2 mm is therefore likely to contribute to the observed increased in soil carbon sequestration in grassland monoliths exposed to low disturbance