What comes next could be worse: Lessons from Hawaiian seasonal forests

Background/Question/Methods

Efforts to control exotic species often assume that the replacement species will be ones with higher ecological value or lower economic damage than those being controlled. Yet rarely are removals conducted experimentally, and successional trajectories are rarely followed after removal. Such studies would inform managers about species likely to dominate in both the short (post control)  and the long term.  Likewise, few studies follow unmanaged sites long enough to ascertain whether invaders may lessen in dominance over decadal time scales without any management at all. We address these issues using long term data and experimental evidence from seasonally dry habitats in Hawaii Volcanoes National Park invaded by perennial bunch grasses, starting with Schizachyrium condensatumin the 1970s.  We compare the trajectory of sites degraded by wildfire and invasion with nearby unburned primary forest sites that have never burned. We ask what the potential role of early invaders was in influencing current vegetation trajectories, and whether vegetation trajectories, and subsequent species invasions could have been predicted.

Results/Conclusions

Our long term data show that although initially invaded by the same grass species, the burned and unburned sites had different successional trajectories after wildfire, as burned sites quickly became dominated by a different species of invasive grass, Melinis minutiflora.  Interestingly, burned and unburned habitats are converging in composition over time (25 years) as yet another invader, the woody N-fixer Morella faya, has invaded the sites.  This invader was rare in the system when the study began in 1990.  Removal experiments conducted in the 1990s showed the possibility that decline of the initial grass invader, Schizachyrium, could enhance invasion rates by Morella in intact woodland sites, yet Schizachyrium decline was not expected since it replaces itself well in unburned sites.  Likewise in burned sites, outplant experiments with native and exotic species demonstrated that the most likely invader was Morella but only when grasses are reduced. We have observed in fact that invasion by Morella has progressed as grass dominance has fluctuated.   We advocate for more experimental work in invaded sites to understand how control of one invader is likely to influence future invaders and discuss the importance of closer working ties between ecologists and managers in planning for future invasion and the outcome of species control efforts.