Changes in soil-surface CO2 efflux following nonnative feral pig (Sus scrofa) removal in Hawaiian tropical wet forest

Background/Question/Methods

Hawaiian native forests are undergoing substantial change due to invasion by nonnative, feral pigs (Sus scrofa).  Feral pigs disturb soil by mixing soil horizons, disrupting soil structure, and decreasing litter cover.  Together, these impacts likely alter soil-surface CO₂ efflux (‘soil respiration’; F_S), the second largest flux in the terrestrial carbon cycle, but few studies have quantified feral pig impacts on soil biogeochemistry.  Here, we quantified F_S inside (pig removal) and outside (pig present) sites along a chronosequence of exclosures ranging in age from 6.5 to 18.5 years since pig removal in native tropical wet forest on the Island of Hawaii.  Our hypotheses were that (i) F_S would be lower in pig present sites due to pig-induced soil disturbance; (ii) the difference in F_S between pig removal and pig present sites would increase with time since pig removal; and (iii) spatial heterogeneity in F_S would be higher in pig present sites.  These hypotheses were tested by quantifying F_S once on twelve 20cm diameter collars in four plots in each of five paired sites (pig removal vs. pig present) across the 18.5 year chronosequence. All sites were located in areas with similar substrate type/age, climatic conditions, and plant community composition.

Results/Conclusions

In line with our original hypothesis, across all sites F_S was ~10% lower in pig present sites (P<0.05), averaging 3.64 and 3.28 mmol CO₂ m^-2 s^-1 in pig removal and pig present sites, respectively.  In addition, the relative difference in F_S between paired sites increased linearly and positively with time since pig removal (r²= 0.80; P<0.01), ranging from a difference of 2% at 6.5 years following pig removal to ~20% at 16.5-18.5 years following pig removal.  Differences in F_S may have been related to a 14% decrease in soil volumetric water content in pig removal sites (P<0.05).  Feral pigs had no clear discernable pattern on the spatial variability in F_S.  Across all sites and treatments, the average number of collars required to estimate F_S within 10 and 20% of the actual mean at a 95% confidence level was 90 and 22, respectively.  These results demonstrate that nonnative, feral pigs strongly influence F_S, with the potential to significantly alter terrestrial carbon cycling in forests impacted by the invasion of nonnative feral pigs, an increasingly common scenario throughout the world.