Early sowing is more effective in the long-term for restoring sandy grassland than six years of mowing or carbon amendment.

In order to find the best and most reliable practices for ecological restoration of degraded lands, longer time scales should be considered when evaluating restoration efforts. We assessed the long-term (16 years) effects of different grassland restoration treatments – seeding, mowing, and carbon amendment – in the Pannonian sand grasslands. After re-plowing in 1 m × 1 m plots, treatments were carried out in two abandoned croplands. Seeding was applied only initially (2002) while mowing and carbon amendment were carried out for six years (2003–2008). Vegetation was surveyed yearly from 2003 to 2008 and re-sampled in 2019 in each permanent treatment plot. We used principal coordinates analysis to describe the trajectories of vegetation development and linear mixed-effects models to test changes in the relative cover of native sand grassland (target) and invasive (neophyte) species with time and treatments. Relative cover of target species increased while neophyte species decreased with time in both sites. There was a higher relative cover of target species from the first or third year on and a lower relative cover of neophyte species from the third year on in one site in seeded plots compared to other treatments. Seeded species also spread into non-seeded plots by 2019, obscuring the differences between treatments 16 years after sowing. Carbon amendment proved to be beneficial in the early and mowing in the later phases of restoration. Based on the long-term results, initial seeding is the best method for restoring sand grasslands in old fields by favoring the establishment of target species and controlling non-native invasion. As a supplement to seeding, carbon amendment can be suggested in the initial phases and/or low-intensity mowing in the later phases of the restoration after land abandonment. Although the spread from seeded plots obscured the long-term differences between treatments, it optimized the restoration process, suggesting that the use of small seed introduction units can be enough to restore the whole degraded area. • Long-term monitoring is needed to better interpret restoration interventions. • Early seeding favors target species and controls invasion in grassland restoration. • Small introduction plots can trigger restoration of larger areas. • Low intensity mowing can be applied in the later phases of the restoration process. • Carbon amendment is suggested in combination with seeding right after abandonment. [ABSTRACT FROM AUTHOR]