RESEARCH ARTICLE Succession in soil seed banks and its implications for restoration of calcareous sand grasslands Péter Török 1,2 , András Kelemen 3 , Orsolya Valkó 3 , Tamás Miglécz 3 , Katalin Tóth 3 , Edina Tóth 4 , Judit Sonkoly 1 , Réka Kiss 4 , Anikó Csecserits 5 , Tamás Rédei 5 , Balázs Deák 3 , Péter Sz ˝ ucs 6 , Nóra Varga 5 , Béla Tóthmérész 3,4 Spontaneous succession is increasingly involved in grassland restoration, because it offers a cost-effective solution compared to technical reclamation methods. This topic is especially important nowadays, as large areas of marginal croplands are being abandoned on poor soils of Central and Eastern Europe, which offers a possibility for the spontaneous recovery of typical target grasslands. Studying the vegetation composition, aboveground biomass, and seed bank in old fields of different age and target calcareous sand grasslands using the chronosequence method, we aimed to answer the following questions: (1) Which species contribute to the seed banks of old fields and reference grasslands? (2) Does the direction of vegetation and seed bank succession trend toward the reference grasslands? (3) How are the vegetation changes in spontaneous succession reflected by the soil seed banks of old fields? In reference grasslands on the dune tops only sporadic seed banks were detected, while several hygrophytes had dense seed banks in reference grasslands in dune slacks. Similarity between the species composition of vegetation and seed banks was low. The development of vegetation and seed banks in old fields progressed toward that of target grasslands and the proportion of weedy species (e.g. indigenous weeds and invasive species) also decreased with time. The cryptogam biomass correlated significantly negatively, while the soil phosphorus significantly positively with the weedy species seed bank density. Our results indicated that the role of persistent seed banks in the regeneration of calcareous sand grasslands from old fields is rather limited and promising vegetation changes are mostly driven by spatial dispersal. Key words: dispersal, grassland restoration, land abandonment, passive restoration, vegetation recovery, weed Implication for Practice • Restoration of calcareous sand grasslands from old fields is weakly supported by the soil seed bank as persistent seeds of characteristic species are mostly lacking both in old fields and in reference grasslands. • Low similarities between vegetation and seed banks stress that promising vegetation changes are most likely driven by spatial dispersal. • Landscape-scale patterns, such as availability and the spatial configuration of reference grasslands, should be evaluated and the support of spatial dispersal should be prioritized. • Fast recovery of cryptogam cover would suppress the establishment of weedy species and might enable the considerable reduction of their seed banks in the long run. Introduction Grassland ecosystems are facing large-scale degradation world- wide. The most crucial drivers of their biodiversity deterioration are area loss and altered management (i.e. cessation of former management or intensification; Bakker & Berendse 1999). In many Western European countries, the area loss of seminatural grasslands was higher than 90% (Diemer et al. 2001; Critchley et al. 2004; Dengler et al. 2014). The area loss is also high in the Western Palaearctic steppe zone; the loss of steppe vegeta- tion is alarmingly high, e.g. in Ukraine, Russia, and Hungary (Deák et al. 2016; Wesche et al. 2016). Grassland fragments in mountain regions are threatened by cessation of former manage- ment (Valkó et al. 2012), while in the lowlands by both cessation and intensification (e.g. high-stocking rates, frequent mowing, over seeding, and/or use of fertilizers; Dengler et al. 2014). As grasslands highly contribute to the landscape-scale biodiversity, their conservation and restoration is an essential task. Author contributions: PT, OV, BD, AK, BT designed the research; all authors contributed to data collection, biomass sorting, and seedling emergence; PT analyzed the data; BT, BD, AK contributed to the analyses; NV, PSZ provided data on cryptogams; PT, OV provided the first draft; all authors revised the manuscript draft. 1 MTA-DE Lendület Functional and Restoration Ecology Research Group, Egyetem Sqr. 1, Debrecen H-4032, Hungary 2 Address correspondence to P. Török, email molinia@gmail.com 3 MTA-DE Biodiversity and Ecosystem Services Research Group, Egyetem Sqr. 1, Debrecen H-4032, Hungary 4 Department of Ecology, University of Debrecen, Egyetem Sqr. 1, Debrecen H-4032, Hungary 5 MTA Centre for Ecological Research, Institute of Ecology and Botany, Alkotmány str. 2–4, Vácrátót H-2163, Hungary 6 Department of Botany and Plant Physiology, Eszterházy Károly University, Leányka Str. 6, Eger H-3300, Hungary © 2017 Society for Ecological Restoration doi: 10.1111/rec.12611 Supporting information at: http://onlinelibrary.wiley.com/doi/10.1111/rec.12611/suppinfo S134 Restoration Ecology Vol. 26, No. S2, pp. S134–S140 June 2018