Resilience, persistence and relationship to standing vegetation in soil seed banks of semi‐arid Australian old fields

Abstract Questions: Do soil seed banks of semi‐arid grasslands reassemble after abandonment from cultivation? Do seeds of native and exotic species persist in the soil? Does time since abandonment affect compositional similarity between the vegetation and seed bank? Does the seed bank contribute to resilience in the vegetation? Location: Native grasslands in northern Victoria, Australia. Methods: Seed bank sampling was conducted in spring and autumn over 3 yrs, across a 100‐yr chronosequence. Species richness, composition and germinant density were determined using the seedling emergence method. Seed persistence was assessed by comparing seed densities in spring and autumn. Seed bank composition was compared with the vegetation. Results: The spring seed bank was dominated at all stages by sedges and rushes; hence, native species richness and seed density were largely unaffected by abandonment. In autumn, grassland species contributed more to the seed bank, but richness was reduced after abandonment and showed little recovery, although seed density partially recovered. Seed bank composition showed some recovery in both seasons. Most species had low persistence in the soil. Compositional similarity between the vegetation and seed bank was greater in old fields than uncultivated grasslands in spring, but not autumn. Conclusions: Resilience varied among seed bank parameters and seed banks had low functional importance. Patterns in the seed bank followed, rather than caused, those in the vegetation. Thus, vegetation recovery cannot rely on the seed bank and persistent seeds were not the key mechanism of resilience in the vegetation.