1. Seed bank and seed rain represent dispersal in time and space. They can be important sources of diversity in the rural landscape, where fragmented habitats are linked by their histories.
2. Seed bank, seed rain and above-ground vegetation were sampled in four habitat types (abandoned semi-natural grassland (ABA), grazed former arable field (FAF), mid-field islet (MFI) and grazed semi-natural grassland (SNG)) in a rural landscape in southern Sweden, to examine whether community patterns can be distinguished at large spatial scales and whether seed bank and seed rain are best explained by current, past or intended future vegetation communities.
3. We counted 54 357 seedlings of 188 species from 1190 seed bank and 797 seed rain samples. Seed bank, seed rain and above-ground vegetation communities differed according to habitat. Several species characteristic of managed grassland vegetation were present in the seed bank, seed rain and vegetation of the other habitats.
4. The seed banks of SNGs and the seed rain of the FAFs were generally better predicted by the surrounding above-ground vegetation than were the other habitat types. The seed rain of the grazed communities was most similar to the vegetation in the FAFs, while the seed banks of the abandoned grasslands most resembled the vegetation in SNGs.
5. Gap availability and seed input could be limiting the colonisation of target species in FAFs, while remnant populations in the seed bank and the presence of grassland specialists in the above-ground vegetation indicate that abandoned grasslands and mid-field islets could be valuable sources of future diversity in the landscape after restoration.
6. Synthesis and applications. SNG communities are able to form seed banks which survive land-use change, but their seed rain does not reflect their above-ground communities. It is important that grassland plants set seed. By connecting existing grasslands with restoration targets, increased disturbance in the target habitats would allow for colonisation via the seed bank or seed rain, while decreased grazing intensity would benefit seed production in the source grasslands. Otherwise, landscape-wide propagule availability might increase with a more varied timing and intensity of management.