Nature will have its way: local vegetation trumps restoration treatments in semi-natural grassland.

Questions How do various grassland restoration methods (hay transfer from local or distant donors, seed sowing or natural regeneration) perform in restoration of species-rich, semi-natural grasslands? Are any of the donors particularly well suited as reference vegetation? Do bare soil treatments give better results than treatments where the soil is covered? What are the probabilities of steering vegetation development towards donor vegetation under the various restoration treatments? Location Sunnfjord open-air museum in rural, Western Norway. Methods In a replicated field trial we tested four restoration methods: transferring fresh hay from a local or a distant donor (hay transfer), seeding with a regional species mix, and allowing natural regeneration. We evaluated outcomes by comparing species richness, transfer rates and vegetation dynamics in donor and trial vegetation over four successive years. Results Our study revealed considerable changes in species composition under all restoration treatments, which decreased over time, indicating decelerating succession rates. Many non-trial grassland species became established, especially after seeding and natural regeneration (bare soil treatments). All trajectories became synchronized from the second year, and by trial end, the restoration treatments converged with respect to species richness and species composition. Conclusion All restoration treatments gave high resemblance to a specific donor: the local one. Distant hay transfer resulted in larger temporal changes and exemplified a climatic donor-receiver mismatch, while the failure of several seeded species to germinate indicated an ecological mismatch under the seeding treatment. Vigorous germination of non-trial seeds in the two bare soil treatments evened out the initial treatment-related differences. Dispersal from the surrounding area and between treatments may be utilized for restoration purposes, but also limits the potential for restoring a specific species composition. Although we apparently failed to tailor donor species composition to the recipient, we note that the slow dynamics of long-lived grassland species may lead to a different long-term restoration outcome. [ABSTRACT FROM AUTHOR]