Mass effects, clonality, and phenology but not seed traits predict species success in colonizing restored grasslands

Preserved grasslands commonly host species-rich plant communities but a large part of the grasslands were plowed up in the past. Their restoration often requires a long time and initial restoration measures might trigger ecosystem recovery, which is then followed by spontaneous colonization. We evaluate the establishment success of target grassland species, which were not sown but established spontaneously in the restored grasslands of Bile Karpaty Mts., SE Czech Republic. According to their key functional traits and incidence in the landscape (mass effect; acquired from the results of a grid mapping project in the region), we examined the frequency of species and their mean cover in 82 restored grasslands. The best predictor of species frequency in the grasslands was their mass effect, followed by a high capacity for clonal growth and late phenology. Seed dispersal traits (seed mass, terminal velocity, epizoochory ranking index) and plant height had no significant effect. Specific leaf area was positively correlated with mass effect. Species having a high cover in the restored grasslands had a high capacity for clonal growth. In the preparation of seed mixtures, we should therefore consider that nonclonal species relying on regeneration from seeds will be generally less able to reproduce and should be promoted by artificial sowing. At the same time, species common in the landscape, which spread well clonally, and those with a late phenology, might be expected to colonize restored meadows on their own, so that sowing them is not necessary.