Separating fast from slow cycling soil organic carbon – A multi-method comparison on land use change sites

Soil organic carbon (SOC) is significantly affected by land use change (LUC). Consequently, LUC is a major controlling factor of total SOC contents and SOC pool dynamics. Several methods have been developed to assess distinct SOC pools, which includes particle size separation, thermal analysis and soil reflectance mid-infrared spectroscopy. All of which are considered to have a potential as high through put methods to generate large datasets. Here, we used 23 sites covering six different types of LUC to assess differences in fast and slow cycling SOC derived from three approaches. We used i) particle size fractionation to obtain coarse (>50 µm) and fine (<50 µm) SOC fractions; ii) thermal Rock-Eval® 6 analysis in compilation with the PARTYSOCv2.0EU model to estimate active and stable SOC pools and iii) mid-infrared spectroscopy to determine the relative SOC composition and derive fast (aliphatic compounds) and slow (aromatic/carboxylic compounds) cycling SOC pools. The particle size SOC fractions and thermal SOC pools showed similar dynamics but differed substantially in the magnitude with LUC. The fine SOC fraction contained around two-thirds of the total SOC across all land uses and was strongly responsive by nearly matching the relative changes of total SOC (slope of 0.76 and R2 = 0.91). Therefore, the fine fraction SOC might be more dynamic than considered until now. In comparison, the stable SOC pool calculated using PARTYSOCv2.0EU was less responsive to the relative changes (slope of 0.43 and R2 = 0.72) and contained around 40 % of the total SOC. This underlines that both physical and thermal approaches separate biogeochemically distinct pools. The qualitative assessment by mid-infrared spectroscopy related well to the thermal SOC pools but not to the particle size fractions. The initial land-use SOC composition, as a ratio of the corresponding fast and slow cycling SOC pool, can be a suitable predictor for SOC evolution. This was particularly true for thermal and mid-infrared spectroscopy derived SOC pools. We show that three conceptually different methods (physical, thermal and mid-infrared spectroscopic) are suitable to determine SOC pool changes for a large diversity of LUC, but the sensitivity of the individual pools can differ strongly, depending on the method.