Removal of non-native trees fosters but alone is insufficient for forest regeneration in Hawaiʻi

Abstract

Across the globe, non-native plant species have become abundant in many tropical forests, resulting in altered patterns of biological diversity and species composition, and impacting important ecosystem functions. However, long-term experimental research on the efficacy of non-native tree removal for tropical forest restoration remains limited. We investigated the removal of non-native tree species in a mesic tropical forest in Hawaiʻi, where rates of endemism and endangerment of plant species is high and multiple non-native tree species are abundant. As a collaboration between resource managers and researchers, we tested the effects of non-native tree removal on overstory and understory metrics of restoration using three restoration approaches: “total cut” (cutting all non-native trees); “girdle” (girdle of all non-native trees); and “selective cut” (cutting approximately 50% of non-native trees). Prior to removal, we established permanent plots (four 10 × 10 m plots per restoration treatment, with 10 1 × 1 m subplots in each), then monitored them over 10 years. Across all treatments, canopy openness increased significantly post restoration, peaked after three years, then decreased to pre-treatment levels or lower. The increase was largest for the total cut treatment, but there was large variation within treatments. By the end of the experiment, the total cut and girdle treatments performed similarly for all our metrics of restoration, including survival, growth, density, and basal area of native trees; density and richness of native species in the understory; and total weed biomass. The selective treatment performed worse, showing lower relative gains in basal area of native trees and density of native understory individuals. Overall, tree removal was effective in restoring a native canopy. However, understory native species richness and density remained at pre-treatment levels, likely due to limited seed dispersal, seed predation by non-native rodents, and continued competition from non-native herbaceous species. Nonetheless, subplots with the best starting conditions performed the best. More intensive and longer-term weeding of non-native species in the understory, focused on areas with the best starting conditions and combined with out-planting of native seeds, seedlings and/or saplings, is likely necessary to foster effective native species regeneration. Overall, our research shows that removal of non-native trees using the total cut and girdle techniques can foster native forest restoration in the mesic forests of Hawaiʻi, but that on-going long-term management in the understory is critical.