Abandoned tropical pastures offer excellent opportunities to restore native forest communities but multiple barriers to restoration often need to be overcome to re-establish native forests. In Hakalau Forest National Wildlife Refuge on Hawai’i island, managers have planted over 400,000 native Acacia koa (koa) trees in abandoned pastures to facilitate forest regeneration. Their goal was to restore canopy trees to promote native understory recruitment by increasing bird-mediated seed rain while decreasing competition from invasive grasses. However, natural recruitment of native understory has not occurred after nearly 30 years. By contrast, recruitment of native understory species does appear to occur under remnant canopy trees of a different native species, ‘ōhi’a (Metrosideros polymorpha). We assessed potential barriers to native plant regeneration in these restoration forests by manipulating thresholds of (1) seed rain and (2) grass competition and evaluating how they interact to influence woody plant establishment. Under individual koa and ‘ōhi’a trees we implemented a factorial experiment that varied grass biomass (100, 50 or 0% removal) and seed addition (0, 10, 20, 50, 100, or 200 seeds/m2) for two common native understory species (Pilo - Coprosma spp, ‘Ōlapa – Cheirodendron trigynum). Seedling establishment was monitored every two weeks and survival tracked through time.
Results/Conclusions
Pilo recruitment was higher under ‘ōhi’a than under koa (p<0.01) suggesting that canopy trees can significantly alter recruitment of the understory. In addition, there was an interaction between grass removal and seed addition on seedling recruitment (p=0.053). Recruitment was elevated in the 50 and 100% grass removal treatments relative to 0% grass removal indicating that grasses suppress native seedling recruitment. However, we observed seedling establishment in the high seed density plots even when grass was not removed, indicating that seed dispersal may also limit native seedling recruitment.
For ‘ōlapa, germination through the first six months of the experiment has been negligible. It is unclear if seeds are exhibiting dormancy or if additional factors are preventing recruitment in the restoration forests (e.g. altered microclimate or missing mycorrhizal associations). These findings suggest that thresholds to natural understory regeneration exist that are holding restoration forests with invasive grass understory in a relatively stable state, making restoration difficult. With no grass removal, seedling establishment occurs only at high levels of seed addition; levels which are an order of magnitude greater than current bird mediated seed rain. Removing just half of the grass cover can increase seedling establishment, at least in the short term.