Challenges in predicting ecosystem response to invasion: A case study of decline of a dominant invader in Hawaii’s seasonally dry woodlands

Background/Question/Methods

The impacts of plant invaders are often measured when they are dominant or co-dominant at a site and then extrapolated to infer long-term ecosystem outcomes. Yet for most invaders, we do not know their longterm persistence and or how their impacts change over time. Longterm sites that we established in invaded seasonally dry tropical woodland sites in Hawai’i offer an opportunity to evaluate how initial impacts translate into longer term persistence and ecosystem change. The dominant non-native species during the initial 30 years of invasion at these sites was Schizachyrium condensatum, a new world perennial bunchgrass which filled in spaces between native species in the woodland understory. Removal experiments conducted in the 1990s showed that this grass reduced growth and recruitment of native woody species while simultaneously reducing invasion rates by two other aggressive non-native species, the n-fixing tree Morella faya and the African grass Melinis minutiflora.  Here, we resurveyed these sites and evaluated changes in dominance and NPP patterns in these woodlands.

Results/Conclusions

Even though S. condensatum produced the most biomass of any single understory species and almost as much biomass in the understory as all other species combined at the height of its invasion, it has not maintained its dominance:  Our recent data demonstrate that S. condensatum has declined to less than a quarter of its previous cover and >75% of S. condensatum individuals in the woodland are dead.  The decline of this initial invader was not predicted in earlier work which had demonstrated its ability to recruit beneath itself and beneath the woodland canopy.  We also found that although native woody species have increased in cover and biomass since initial measurements, simultaneous with the decline of S. condensatum, the other two more aggressive invaders, M. faya and M. minutiflora, have increased with potentially more dramatic consequences for the native woodland species. While the mechanism of S. condensatum decline is not yet clear, it highlights the need for long-term data in invasion biology.   Future work comparing how and why invasive species can, in the long-term, lead to persistent invader dominated states versus simply changing successional dynamics will help ecologists predict long-term ecosystem trajectories.