Active seed sowing can overcome constraints to passive restoration of a critically endangered vegetation type

Abstract Invasive alien plants negatively impact ecosystems, but recovery of native vegetation may fail following standard methods of alien species removal alone. Alternative management actions may thus be required. Cape Flats Sand Fynbos is a critically endangered vegetation type in the Cape Floristic Region of South Africa which is threatened by Acacia saligna invasion, but standard clearing methods have failed to restore native vegetation structure. A restoration study was performed comparing passive treatments i.e. clearing without burning (stack-block) versus clearing and burning (burn-block), as well as active intervention by sowing seeds of native species, either initially after burning or a year later, in which seeds were either not pre-treated or pre-treated with smoke and heat exposure before sowing. After two years all treatments resulted in different recovery trajectories, although none resembled the reference condition. Clearing without burning facilitated recovery in less degraded areas with higher initial native shrub cover, but otherwise resulted in limited vegetation recovery. Limited recovery facilitated secondary invasion by herbaceous weeds. Active seed sowing resulted in the highest recovery of native shrub cover and diversity. These findings suggest that passive restoration is constrained by seed limitation, due to the lack of recovery of vegetation components under passive clearing treatment. Active sowing was able to partially overcome this constraint through improved recovery of total shrub cover. However, non-sprouting shrub cover was higher while resprouting shrubs and species of Restionaceae were lower compared to the reference condition. Pre-treatment of seeds before sowing improved establishment of some species. Active treatment involving sowing pre-treated seeds after clearing and burning therefore resulted in best fynbos recovery compared to either of the passive treatments tested. A decision tree has been developed based on these findings in order to guide best protocol for managers.