Reproducibility of a soil organic carbon fractionation method to derive RothC carbon pools

Fractionation of soil is undertaken to isolate organic carbon with distinct functional properties, such as stability and turnover times. Soil organic carbon (SOC) fractionation helps us to understand better the response of SOC to changes in land use, management or climate. However, fractionation procedures are often poorly defined and there is little information available on their reproducibility in different laboratories. In a ring trial, we assessed the reproducibility of a SOC fractionation method introduced by Zimmermann et al. (2007). The isolated fractions were linked to the model pool sizes of the Rothamsted carbon model (RothC). We found significant differences between six laboratories for all five defined fractions in three different soils with coefficients of variation ranging from 14 to 138%. During ultrasonic dispersion, the output power (energy per unit time) was identified as an important factor controlling the distribution of SOC within these five fractions, while commonly only the output energy is standardized. The amount of water used to wet-sieve dispersed soil slurry significantly influenced the amount of extracted dissolved organic carbon (DOC). We therefore suggest using a fixed amount of power for ultrasonic dispersion (20W) and a minimum amount of water for wet sieving (2000ml). RothC pool sizes were predicted from the measured fractions and compared with RothC equilibrium pool size distributions. This model initialization using measured SOC fractions, however, led to an over-estimation of stable RothC SOC pools when compared with pool size distributions derived from RothC equilibrium runs under a bare fallow soil model simulation. To improve the isolation of particulate organic matter from stable mineral-bound organic matter, we suggest that the density should be increased from 1.8 to 2.0gcm(-3) in the density fractionation step. We formulated a modified fractionation procedure, which aims specifically to enhance reproducibility across laboratories and to improve the match of the isolated SOC fractions with RothC's SOC pools.