Rapidly aging American dam infrastructure has motivated calls for dam removal as a way to cut costs, improve safety, restore riparian habitats and improve fish passage. However, there is little data on the effects of dam removal and almost no long term data on the effects of removing dams taller than 5 m. The Elwha River dam removal (Washington, USA), the largest to date, was completed on the downstream (Mills) reservoir in 2012, and upstream (Aldwell) in 2014, exposing a cumulative 2.76 km2 of previously-inundated sediment surface. The revegetation of reservoir beds on the Elwha is a unique opportunity to study vegetation restoration following dam removal. Our objective was to determine how different reservoirs and landform surfaces within the reservoirs differed in terms of riparian vegetation succession in the absence of active restoration. We surveyed vascular plant species richness and cover in 65 100m2 permanent plots in 2013, 2014, and 2016. The plots are located on 10 transects perpendicular to river flow and randomly stratified across reservoir landforms (terrace, valley wall, and active channel). We compared changes in native and nonnative species richness and cover over time across the different landforms and reservoirs using mixed linear models.
Changes in species richness and cover from 2013 to 2016 varied across reservoirs and landforms. Species richness declined on the valley walls (p=0.005) and terraces (p=0.002) of Aldwell Reservoir; while cover of exotic species increased significantly on these landforms. Concurrently, species richness increased in the active channels of Aldwell Reservoir, probably due to the combination of frequent disturbances and seed rain. This is in line with predictions about successional stage of the landforms. In Mills Reservoir terraces, native species richness has increased rapidly compared to terraces in Aldwell Reservoir (p<0.001). Mills Reservoir plant communities have had less time to develop, so we expect future conditions to become similar to current Aldwell conditions. In 2016, non-native species cover was the same across all landforms and in both reservoirs. Because the Elwha dam removals are recent, rapid changes in vegetation are to be expected, however it is unclear the degree to which forest communities that develop will reflect pre-dam vegetation. In many cases, such as terrace surfaces where large amounts of remnant reservoir sediments have accumulated, novel plant communities may develop. Continued monitoring of this project will be essential for understanding the recovery of native vegetation following dam removal.