This experiment has achieved the goal of slowing down nutrient cycling in the forest; nutrient cycling is at least 80% lower, regardless of treatment! Moderate redundant plots have shown the most combined survival and growth of outplants. Redundant plots have the greatest native seedling recruitment. Moderate complementary plots appear to be the least resistant against invasive species.
ʻIeʻie (Freycinetia arborea) is only typically found in lowland wet forests, and is an indicator of a healthy forest. We have found ʻieʻie growing in many of our plots!
We often see ʻIo (Buteo solitarius), the Hawaiian hawk, in the experimental plots!
Survival and Growth
Slow redundant plots have the greatest outplant survival over the years, with moderate redundant in second (Figure 9). However, slow redundant sees the lowest growth based on basal area, while moderate redundant plots have the greatest (Figure 10). Some outplants that did especially well in terms of survival and growth include kukui, ʻulu, mango, false kamani, neneleau, avocado, kamani, and ʻōhiʻa ʻai. On the other hand, māmaki, hame, and kōlea lau nui had very low survival.
Figure 9. Outplant survival (%) across the treatments for 6 census dates since the experimental start in January 2014. Values are means + SE. Letters that are different at the last timepoint in February 2018 indicate significant differences (p ≤ 0.05) among treatments.
Figure 10. Outplant basal area across the treatments for the census date (February 2018) that represents five years of growth. Values are backtransformed least-square means +/- 95% confidence limits. Letters that are different denote statistically significant differences among the four experimental treatments.
Reproduction and Regeneration
Slow complementary plots show the greatest amount of fruiting, and moderate redundant had the second most fruiting. For all plots but moderate complementary, existing native plants have significantly more seedlings than outplants (Figure 11a). Existing native recruitment was greatest form slow redundant and moderate redundant, and lowest for moderate complimentary (Figure 11b).
Figure 11. Mean number of fruiting outplant individuals per month compared between treatments and across five years of data collection (a). Values are means + SE. Lower case letters denote statistically significant differences among the four experimental treatments for the most recent census year. Seedling recruitment (#/m²) of native seedlings from ‘Existing’ native seed source (filled circles) and planted native/exotic seedlings from the ‘Outplant’ native/exotic seed source (open circles) recorded in 2018, across the treatments (b). Values are least-square means + 95% confidence limits. Letters denote statistically significant differences across the four experimental treatments.
Invasion Resistance
Weeding effort is greater in moderate complementary plots than all other treatments (Figure 12), suggesting it is the least resistant to invasive species encroachment.
Figure 12. Weeding effort (# person hours per plot), for the most recent 6-month maintenance interval (July 2018-Dec 2018), which represents the fifth year of the experiment. Values are backtransformed least-square means +/- 95% confidence limits. Letters denote statistically significant differences among the four experimental treatments.
Nutrient Cycling
Different functional traits modify the environment in different ways. One goal of this project was to reduce light levels in the understory as well as the rate of nutrient cycling to make the forest lest prone to invasion. A greater leaf area index equates to more leaf layers in the canopy and therefore a darker understory. Moderate redundant plots had the greatest leaf area index, whereas slow complimentary and moderate complementary plots had the lowest (Figure 13). Moderate redundant plots also had the greatest amount of nutrients released from litterfall, significantly more than slow redundant and slow complimentary plots (Figure 14).
Figure 13. Leaf area index (LAI, m^2 leaf m^-2 ground) in the four treatments. Values are backtransformed least-square means + 95% confidence limits. Letters denote statistically significant differences among the four experimental treatments.
Figure 14. Nitrogen (N) released from litterfall across the four experimental community treatments (SLOW RED, MOD RED, SLOW COMP, and MOD COMP). Values are back-transformed means +/-95% confidence intervals. Letters indicate significant differences among treatments. The same patterns were seen for two other nutrients: carbon (C) and phosphorus (P).