Title: Patterns and consequences of re-invasion into a Hawaiian dry forest restoration
Abstract: The restoration of native plant diversity may be an effective tool for weed control, but its use has not been tested in the heavily invaded Hawaiian dry forest ecosystem. In addition, the ecological mechanisms by which invasive plants may cause declines in native plant diversity are generally not well understood. We examined invasion resistance and the relationships between invasion and declines in native plant diversity at local scales in a Hawaiian dry forest restoration by experimentally removing non-native species, planting native species, manipulating resources and environmental conditions, and allowing non-native species to re-invade. We found higher invasion rates in habitats that supported the most native species, suggesting that similar mechanisms may regulate the distribution of both native and invasive species. Pennisetum setaceum, a dominant invasive perennial grass, was associated with native plant mortality and declines in native diversity. Although invasion rates were greatest in more favorable habitats, the association of invasion with native species loss was significant only in lower quality habitats, suggesting that environmental conditions may regulate competitive interactions between native and invasive species. We found that native woody plant restoration is not an effective tool for weed control in this community, and that invasion may result in declines in native diversity. Our study provides a caveat to previous suggestions that invasion impacts on native diversity will be greatest in the most diverse plant communities. At local scales in some communities it may also be effective to control invasive species in sensitive areas with low resource availability instead of, or in addition to, areas with high biodiversity. [Erin J. Questad, Jarrod M. Thaxton and Susan Cordell (2012). Patterns and consequences of re-invasion into a Hawaiian dry forest restoration. Online First™, 6 June 2012 Biol Invasions DOI 10.1007/s10530-012-0253-6] Comment
