Vegetation recovery is site specific following removal of nonnative feral pigs from Hawaiian tropical montane wet forest

Background/Question/Methods

Nonnative feral pigs (Sus scrofa; wild boar) are an exotic disturbance agent that alters native ecosystems on six continents and many oceanic islands.  Feral pigs degrade native vegetation and facilitate invasion by nonnative plants via browsing, rooting, and trampling. Hawaiian forests are thought to be particularly vulnerable to feral pigs, as native vegetation evolved without large mammals. Fencing and eradication of feral pigs is considered a crucial first step in conserving and restoring Hawaiian wet forests. There is little information, however, on the extent and rate of native plant community recovery following feral pig removal, particularly across a range of sites and at regional scales. Because eradication is quite controversial in Hawaiʻi, information on the response of native plant communities to feral pig removal is needed for land management and policy decisions. We examined understory vegetation in paired sites (pig-removal vs. pig-present) inside and outside of five ungulate exclosures representing a chronosequence of feral pig removal (6.5 to 18.5 years) in a ~40km² area of wet forest on the Island of Hawaiʻi.  We quantified forest floor cover, and richness, abundance and rooting location (epiphytic vs. mineral soil) of understory plants in thirty six 1000 m² plots.

Results/Conclusions

Removal of feral pigs substantially impacted understory native vegetation. Native plants were significantly more abundant in pig-removal sites, and this was true regardless of amount of time since feral pig removal. Cover of native herbaceous vegetation was 2x higher (P<0.004) and stem density of species of conservation concern was 4x higher (P<0.019) in pig-removal sites. Feral pig removal also affected recruitment patterns. Epiphytic establishment was similar with and without feral pigs, but density of ground-rooted native woody plants was 4x higher in pig-removal sites (P<0.005). Epiphytic and ground-rooted plant community composition also differed, suggesting long-term impacts of feral pigs on forest structure. Species richness of ground-rooted native woody species was greater in pig-removal sites (P<0.048). Area of exposed soil was extensive where feral pigs were present (22.6%) and minimal in pig-removal sites (1.7%). Litter cover was 67.1% and 50.6% and bryophyte cover was 7.6% and 4.4% in pig-removal and pig-present sites, respectively. Across the chronosequence, our data revealed highly site-specific trends in vegetation recovery following feral pig removal, implying that management and restoration should be based on local site conditions and take into account extent of degradation, presence of nonnative invasive plants, and initial native floristic composition.