RESEARCH ARTICLE Near-natural methods promote restoration of species-rich grassland vegetation—revisiting a road verge trial after 9 years Inger Auestad 1,2 , Knut Rydgren 1 , Ingvild Austad 1 The present loss of species-rich grasslands makes it vital to restore these valuable habitat types, including novel habitat variants such as road verges. Due to the lack of knowledge on long-term outcomes of restoration initiatives, well-designed studies comparing different restoration methods are needed. In this study, we examined ine-scale vegetation recovery patterns over 9 years in a ield experiment with several near-natural restoration methods (adding local seed mixtures, transferring hay from local grasslands using hard or light raking, and natural regeneration) in a road verge. We compared this to standard revegetation (hydroseeding species-poor commercial seed mixtures). We found major temporal changes in vegetation restored by local seed or hay transfer, before it gradually became more similar to the donor grasslands and seed mixtures, which served as references for the experiment. Natural (spontaneous) regeneration with seed dispersal from surroundings gave similar results, whereas areas revegetated using standard methods became more dissimilar to the reference sites during the study period. The main variation in species composition relected the contrast between local donor grasslands and seed mixtures and the species-poor early successional grasslands. We conclude that near-natural methods (hay transfer and seeding) successfully restored species-rich grassland, including road verges. This study underlines the importance of comparing several treatments over a suficiently long period to assess their success in restoring species-rich grassland. Key words: grassland, hay transfer, hydroseeding, resilience, restoration, seeding, species richness, vegetation Implications for Practice • Near-natural restoration methods successfully restore species-rich grasslands, but vegetation will change over an extended period before the reference state is approached. • Seeding species-rich seed mixtures transfers a greater fraction of grassland target species than other methods of species introduction, but species-rich local plant material can also be successfully transferred by hay transfer; the litter layer added through hay transfer gives particularly valuable protection in a dry climate. • Natural regeneration depends on seed dispersal from the surroundings and can lead to successful but slower regen- eration than other near-natural restoration treatments. • Standard revegetation methods, including fertilizing and hydroseeding of species-poor commercial seed mixtures, give poor ecological performance, hinder establishment of native species and should only be used where erosion risk is high. Introduction Construction of road verges is one of the many types of interventions that regularly disturb ecosystems and remove vegetation, thus acting as reset mechanisms for succession (Glenn-Lewin & Van Der Maarel 1992). There are many kinds of disturbances, and depending on their intensity, timing, and duration (Miller et al. 2011), the vegetation may resist the disturbance, change but return almost completely to the initial state, or change irreversibly (Connell & Slatyer 1977). Insight into long-term vegetation responses after disturbance is a key factor for understanding ecosystem dynamics relevant to restoration (Standish et al. 2014). Although road verges are often regarded as disturbed environ- ments without nature value, vegetation resembling traditional seminatural grasslands has been reported in road verges that are regularly managed for trafic safety reasons (Parr & Way 1988; Auestad et al. 2011). Grassland restoration has become a hot topic after the great decline in the area of seminatural grasslands in most parts of Europe in recent decades (Bakker & Berendse 1999; Poschlod & Wallisdevries 2002). It has been shown, however, that species-rich grasslands can be restored after road Author contributions: KR, IdA conceived and designed the research; IrA carried out the ield work; KR analyzed the data; IrA, KR, IdA wrote and edited the manuscript. 1 Faculty of Engineering and Science, Sogn og Fjordane University College, PO Box 133, N-6851 Sogndal, Norway 2 Address correspondence to I. Auestad, email inger.auestad@hisf.no © 2015 Society for Ecological Restoration doi: 10.1111/rec.12319 Supporting information at: http://onlinelibrary.wiley.com/doi/10.1111/rec.12319/suppinfo Restoration Ecology 1