Quantification, spatial distribution and persistence of root-derived carbon for 12 cover crops

Abstract. Organic carbon derived from roots is a major input fuelling soil organic carbon stocks, especially in agricultural systems, where aboveground biomass might be harvested. However, root sampling for carbon monitoring excludes net rhizodeposition, i.e. the organic compounds released by the roots that have not been rapidly mineralised, and some of the finest root debris, because this fraction of root-derived carbon can not be directly quantified in the field. To compensate for this shortfall, we set up a two-month experiment with multi-pulse ¹³C-CO₂ labellings of 12 crops to quantify these carbon pools at harvest, operationally grouped under the term SOC_new. We also investigated the spatial distribution of belowground carbon inputs within the soil profile. Lastly, in order to follow the fate of this carbon after the plant death, we performed a 524-day litterbag incubation in the field using the labelled material. We found that SOC_new accounted for 27 % of belowground carbon inputs at harvest. It was not correlated to carbon amounts of the shoots, but was positively correlated to root carbon (R² = 0.14). The vertical distribution of SOC_new tended to follow the one of roots. The majority was recovered in the bulk soil, rather than adhering to the roots. We showed that SOC_new had a greater persistence time in the soils than roots in the mid-term. However, these findings were marked by high variability because the small quantities of carbon involved make it difficult to assess persistence by isotopic difference. These results suggest that net rhizodeposition and fine root debris should be taken into account in organic carbon management of soils as it refines our estimation of belowground inputs. However, their low predictability, due to the diversity of products and processes, is still a barrier.