Seedbank-Vegetation Relationships in Restored and Degraded Annual California Grasslands: Implications for Restoration

Abstract

Seedbanks represent the pool of species that could germinate and establish into plant communities. Seedbank analysis complements vegetation monitoring and has direct applications for restoration. For most restoration projects, viable seeds of planted species must be present in the seedbank for these species to persist. Degraded California grasslands are increasingly restored to reverse widespread loss of these iconic habitats that support numerous species and provide critical ecosystem services. We analyzed seedbank composition and seedbank-vegetation relationships in paired restored and unrestored Central Valley grasslands to determine: 1) if seeded grasses had persisted in seedbanks and/or vegetation; 2) if native grasses were still present in seedbanks and/or vegetation of unrestored sites; and 3) similarity between seedbanks and standing vegetation. We found that few native grasses persisted in seedbanks of restored sites, with only marginally better results in terms of standing vegetation. No native grasses were found in unrestored site seedbanks and only trace amounts were detected in standing vegetation. We also found low similarity among seedbanks and between seedbanks and vegetation at all sites, suggesting a high degree of patchiness and low predictability of seedbank composition from vegetation. These results suggest that some restoration sites in the region are on undesirable trajectories, likely due to re-invasion of nonnative species. Our results further suggest that passive restoration is not a viable approach for restoring degraded sites in the region. We propose that managers incorporate seedbank analysis into monitoring efforts to more fully understand the present and future trajectories of restored and unrestored sites.