Nitrogen-fixing trees are often planted for restoration because they grow quickly, and readily create forest canopy structure. However, by contributing nitrogen to the soil, nitrogen-fixing trees may have unintended side-effects that alter the communities they are intended to restore. Outcomes of these interactions can vary with environmental conditions, yet this is rarely studied. To determine how plant-soil interactions vary with microclimate (represented by stem density) and climate (rainfall), we studied Koa (Acacia koa), a relatively fast-growing, native, nitrogen-fixing tree, frequently planted on the Hawaiian Islands to reforest pasture. Prior work in a single site has found that despite rapidly forming closed canopies, Koa does not consistently suppress pasture grasses. This allows persistent grasses to compete with desirable native forest species attempting to recruit under Koa. We sought to disentangle the plant-soil interactions leading to suppression (or perhaps facilitation) of grasses under different environmental conditions and thereby determine when and where Koa is most effective as a restoration tool. To do this, we examined Koa-soil-grass interactions across a range of stem densities and rainfall on Hawai‘i Island. We measured soil nitrogen, soil moisture, litterfall, Koa litter decomposition, standing pools of litter and grasses, light, percent cover, and composition of understory communities.
Results/Conclusions
We found that less light reaches the forest floor (p<0.05), more leaf litter falls (p<0.05), and less grass is present (p<0.05) with more Koa stems. These patterns demonstrate the potential for Koa to suppress grasses via shading and accumulation of a physical leaf litter barrier. However full grass suppression never occurred in our plots. One possible explanation is that light levels rarely dipped below 5% full sun—a threshold under which grass growth is reduced. A second reason may relate to faster decomposition of Koa litter at high stem densities (p<0.05). This suggests that even when large volumes of Koa leaves fall and could form a layer to suppress grasses, litter does not accumulate. These faster decomposition rates may be fueled by higher concentrations of soil nitrogen (contributed by Koa) where there are more Koa stems (p<0.05). The rainfall gradient had little bearing on these patterns. Overall, our results suggest that planting Koa leads to the creation of a Koa-grass state, rather than reducing competition with pasture grasses to allow native forest species to recruit. Thus, in this system, despite the variation of plant-soil interactions with microclimate created by a range of Koa stem densities, the restoration outcome was unaffected.