Embankment as a carbon sink : a study on carbon sequestration pathways and mechanisms in topsoil and exposed subsoil ; Les talus comme puits de carbone : une étude sur les voies et mécanismes de séquestration du carbone dans le sol superficiel et le sous-sol exposé
Carbon (C) sequestration is receiving increasing scientific and political attention in a framework of greenhouse gasses mitigation. However, geotechnical soils have been neglected for their C sequestration potential, with the global attention focusing on agricultural and natural soils. In the present thesis project we aim to assess the potential of geotechnical embankments as C sink, and, through the study of plant species and soils showing contrasting features, shed light on SOC sequestration mechanisms and the role of the different actor involved. We aim not only to quantify the C gained and lost in soil, but even its origin (fresh new C input or old preexistent C) and how it is partitioned in different C pools characterized by different C stability (quality of stored C). First, we evaluated the C storage in different pools under soil sowed with 12 different herbaceous species in a 10 months experiment. Assessing different root traits allowed understanding the influence of root economic spectrum on C storage. We showed how traits linked to high labile C are linked to a higher C increase in the stable SILT+CLAY pool (<20µm). Root traits related to a low input of recalcitrant, instead, favor accumulation in the unstable POM fraction. Thanks to a 183 days stable isotope labelling experiment (CO2 constantly enriched with 13C) we were able to study the C dynamics in different C pools under two species (L. perenne and M. sativa) sowed on two soil (topsoil, 0-30cm depth and subsoil brought to the surface, 110-140 cm depth) showing contrasting characteristics. We evidenced the great interest of bridging C origin and C pools when studying soil C fates, allowing unveiling processes those more traditional methods would hide. New C and old C showed synergetic covariation, with lower old C losses associated to higher new C inputs. This is in good accordance with the Preferential Substrate Utilization hypothesis (Cheng and Kuzyakov, 2005). The Preferential Substrate Utilization hypothesis was also validated with the ...