Long-term persistence of seeded grass species: an unwanted side effect of ecological restoration.

Spoil heaps are the visible footprint of hydropower production, particularly in vulnerable alpine environments. Speeding up vegetation development by seeding commercial grass species has been a common restoration practice for the last 50 years, but we lack information on whether seeded species decline and allow native plant cover to develop. We visually estimated cover of native vascular plants and five seeded grass species (Agrostis capillaris, Festuca ovina, Festuca rubra, Schedonorus pratensis and Phleum pratense) on eight spoil heaps at different elevations (boreal-alpine zone) in western Norway. Spoil heap vegetation was censused twice (9-20 and 24-36 years after spoil heap construction); the undisturbed surrounding vegetation was also censused on the second occasion. Total cover on the spoil heaps showed some increase, but remained far below that in surrounding areas. Cover of seeded grass species in the surroundings was low (but not negligible), indicating suboptimal establishment ability. Seeded species usually covered less than 20 % of the spoil heaps, and only F. rubra, F. ovina and A. capillaris contributed substantially. Proportional cover indicated better initial establishment by seeded species, but their cover decreased between the censuses on all but the highest located spoil heap. The persistence of seeded grass species is problematic, and despite the decrease in proportional cover, they are likely to persist for decades on spoil heaps, posing a risk of invasion of surrounding areas. We therefore recommend replacing the practice of seeding with more appropriate restoration measures.