Abstract
Worldwide, riparian forests have been impacted by regulation and artificialization processes. Restoration of riparian habitats has yet become a major issue. However, given the alteration of natural disturbance regimes and modifications in abiotic conditions induced by stabilization structures, the efficiency of these operations is questionable. Indeed, communities may have already been oriented toward transient or new states, making the return to the desired state difficult or impossible. To evaluate the potential added-value of restoration operations, it is thus of great concern to compare current trajectories of degraded rivers to historical trajectories of reference rivers. Using a modelling framework, we assessed whether ecological and functional trajectories varied with forest age between the Rhône (constrained, n plot = 65) and the Drôme Rivers (unconstrained, n plot = 69). Results showed that the shift in functional traits (SLA, wood density) was accelerated along the Rhône River and that patterns of trait divergence (wood density, seed mass) with forest age were more pronounced along the Drôme River. For stand attributes, basal area and mean diameter of live and dead trees increased with forest age along the Drôme River but decreased along the Rhône River, while basal area of exotic tree species increased only with forest age along the Rhône River. Furthermore, forest composition and structure strongly diverged between the two rivers. This was mostly due to the greatest abundance of exotic species and of large diameter trees within the riparian forests of the Rhône River. Overall, we found profound differences in ecological successions between the Rhône and the Drôme rivers, highlighting rapid and irreversible divergences in trajectories. Along the highly artificialized margins of the Rhône River, riparian forests could be regarded as "novel ecosystem" for which restoration thresholds are exceeded but where beneficial outcomes remain available (e.g. ecosystem services, biodiversity conservation...).