Seedling recruitment responses to interventions in seed-based ecological restoration of Peninsula Shale Renosterveld, Cape Town

• Seeding was imperative to restoration success.
• Burning is recommended for the positive effects it has on existing indigenous community and the inhibitive effect on weeds.
• In old fields, tillage is recommended for the initial creation of safe-sites but requires follow-up herbicide application.
• Herbicide is recommended for reducing weeds and the positive responses exhibited by seeded and existing indigenous species.
• An ecological-response model provided a framework for considering findings in relation to expected ecological factors.

Peninsula Shale Renosterveld, a Critically Endangered ecosystem, requires ecological restoration intervention on transformed areas. Yet there is little guidance due to very few renosterveld studies and limited knowledge of the mechanisms driving ecological responses. This study set out to test the effects of 32 interventions, comprised of five crossed factors (seeding of 31 species, fire, tillage, herbicide application and rodent exclusion), on vegetation recruitment on a site dominated by alien, annual grasses. Actual experimental responses were compared with predicted responses presented in an a priori ecological-response model. Responses to the interventions were highly variable. Sowing on its own was almost ineffectual, but restoration of species was enhanced when seeding was implemented with one or more of the other factors. In combination with other treatments, seeding made significant contributions to overall seedling density, species richness and canopy cover and is imperative if this ecosystem is to recover from extensive alien grass invasion. Several three- to five-factor interventions resulted in a full set of desired responses. Half of the responses predicted in an ecological-response model were confirmed. Study outcomes have the potential to guide future research and implementation of larger scale renosterveld restoration.