Are Pre-Restoration Soil Seed Banks and Vegetation Nested and Predictive Subsets of Post-Restoration Communities?

ABSTRACT Ecological restoration projects could benefit from using knowledge of pre-restoration conditions to forecast potential restoration outcomes, including species compositional change that can be among the most stochastic changes in ecosystems. In Pinus ponderosa (ponderosa pine) forests in Arizona, USA, predictability of understory plant composition from pre-restoration soil seed banks, on-site vegetation, and nearby vegetation was examined for 12 years after restorative forest thinning treatments. Pre-restoration seed banks and vegetation were nested and predictive subsets of post-restoration species composition of understory communities. For example, the portion of species in the pre-restoration seed bank also in the on-site vegetation increased from 32% before to 79% five years after restoration at nine sites. As many as 57–69% of species only inhabiting seed banks before restoration transitioned to occurring in vegetation, strengthening seed bank:vegetation correspondence. In total, knowledge of pre-restoration seed bank and vegetation composition enabled forecasting 64–76% of the species in vegetation up to 12 years after restoration, with another 14–26% of species predictable from species composition in nearby remnant openings. Ecosystems in which practitioners may anticipate predictability of species compositional changes after restoration could include habitats with high potential for seed bank:vegetation synchrony (i.e. seed banks containing species capable of growing in a site’s vegetation), moderately shaded structure enabling species varying in shade tolerance to coexist, and with seed dispersal (at least stochastic long-distance dispersal) subordinate to local regeneration from seed banks and vegetative expansion. Changes in species composition may be more deterministic than stochastic in at least some ecosystems undergoing restoration.