Wetland restoration: Predicting vegetation trajectories over 25 years

Abstract Worldwide wetland loss has made the conservation of these ecosystems a policy priority and led to the multiplication of restoration programmes. However, the lack of long‐term monitoring limits our understanding of the processes influencing the vegetation composition of restored wetlands and our ability to predict outcomes over multiple decades. Here, we assessed the extent to which hydrological regime and planting density of target species, two critical factors driving wetland vegetation and restoration success, can predict restoration outcomes. Using correlation analyses and generalised models, we assessed the role of target species planting density and analogous hydrological conditions (e.g. level, variation, seasonality) to reference wetlands for achieving and predicting restored vegetation similarity to reference plant communities in 12 sedge and/or willow‐dominated wetlands in Mountain Village, Colorado over 25 years post‐restoration. We found a significant positive correlation between hydrological similarity and vegetation similarity, peaking at 15 years post‐restoration (rho = 0.61). Similarly, planting density was positively correlated with vegetation similarity, peaking 5 years after restoration (rho = 0.75). For both variables, communities with the shallowest water table exhibited the strongest correlations. The similarity of restored vegetation to the reference community can be predicted using hydrological similarity and planting density. The models that combined these two variables outperformed single‐variable models. However, the model accuracy decreased 25 years after restoration, making predictions over two decades inaccurate for most communities. Synthesis and applications. Hydrological similarity to a reference, combined with appropriate planting densities, reliably predicts restored wetland vegetation convergence towards reference communities over two decades. Such models could provide managers with tools to assess failure risks across potential restoration sites, allowing them to select the most suitable locations and tailor planting efforts to maximise wetland restoration success.