In the French Southern Alps, severe erosion in marly badlands shows negative consequencesfor human populations, who worked to restore these terrains since the XIXth century. For adecade, an ecological engineering strategy is developed and focuses on gully beds, cornerstonesof their restoration. The operational question at the root of this PhD project is to identify towhat extent ecological engineering structures implanted in marly gully beds can foster theirlong-term ecological restoration. The scientific objective related to this question is to betterunderstand how plant communities interact in the long-term with the restoration of soils ingully beds. We hypothesized that plant traits at the community scale is a relevant entry point tostudy these interactions. In the first part, we used traits as a tool to study the mechanical effectof plants on geomorphological dynamics in gully beds under restoration. In particular, weshowed that traits enable us to explain the capacity of plant communities to favor theaccumulation of sediment mounds in gully beds during the first decade after ecologicalengineering works. In the second part, we concentrated on the potential long-term (century)evolution of the properties of these sediment mounds, in the process of becoming real soils. Weshowed that both trait diversity and dominant values of traits interacted with the properties ofthese soils, such as soil aggregate stability and soil fertility. With a strong multidisciplinaryapproach, this work provides results and insights on the interactions between plant traits andecosystem functioning in the specific case of highly eroded semi-natural ecosystems. It alsocontribute to set tools to restore badlands via ecological engineering works, by providinginformation about the long-term geomorphological and ecological evolution of soil-plantsystems in restored gully beds.
