Restoration potential of the persistent soil seed bank in successional calcareous (alvar) grasslands in <scp>E</scp>stonia

Abstract

Abstract Questions What is the overall restoration potential of the persistent soil seed bank of abandoned alvar grassland sites? To what degree does the share of characteristic alvar grassland species in the soil seed bank change during secondary succession from traditionally managed pasture towards stages overgrown by C orylus avellana and J uniperus communis ? Location Northern E stonia. Methods The species composition of above‐ground vegetation and the persistent soil seed bank were investigated in successional alvar grasslands. Particular attention was given to typical grassland species in the persistent soil seed bank. Soil seed bank density was also recorded. General linear models and non‐metric multidimensional scaling were used to assess successional trends and the restoration potential of the soil seed bank. Results The cessation of traditional management has led to a considerable decline in above‐ground plant species richness. The composition of above‐ground vegetation in different successional stages was considerably more variable than that of the seed bank. We detected large and species‐rich persistent soil seed banks in all our study sites. About 70% of species in the seed banks of overgrown sites could be classified as typical grassland species. The frequency of typical alvar grassland species, such as A rabis hirsuta, A renaria serpyllifolia , C erastium fontanum , H elianthemum nummularium , P oa angustifolia, P otentilla crantzii , S edum acre and V eronica spicata in the seed bank of long‐term abandoned (25 and 50 yr) sites was unexpectedly high. Conclusions Abandoned semi‐natural grassland sites in northern E stonia still contain a relatively large grassland community species pool, including a large and effective soil seed bank. The soil seed bank could thus play a more important role in the restoration management of abandoned dry grassland communities than could be concluded from previous studies.