Nonnative ungulates exert a large negative effect on native biodiversity and the structure and function of terrestrial ecosystems throughout the Pacific Island region. In Hawaii, removal of ungulates is broadly recognized as a critical first step in conserving native ecosystems and species, particularly threatened, endangered, and at-risk species (TER-S). To this end, land managers commonly fence and remove nonnative ungulates where conservation of native biodiversity is a priority. However, these actions are labor and cost intensive, and the long-term outcomes are not well quantified. Surprisingly little information exists on the magnitude and time frame of native plant recovery across different ecosystem types, the potential for nonnative plant invasions, or the response of critical, underlying ecological processes to nonnative ungulate removal. This study quantified the impacts of nonnative ungulate removal on the biodiversity, structure, and function of two major ecosystem types – tropical wet forest (TWF) and tropical dry forest (TDF) – found on Department of Defense (DoD) installations throughout the Pacific Island region. Specifically, the research team examined pathways and mechanisms through which ungulate removal impacts native and nonnative plant dynamics in four distinct ecosystem/plant community types. They also measured how ungulate removal affects key underlying ecological processes in three of these ecosystem types, focusing on ecosystem carbon (C) storage and flux, and soil nitrogen (N) cycling and availability. They then tested if manipulation of soil nutrients can be used as a management tool to favor native plants over nonnative, invasive plants in both greenhouse and field experiments. Finally, they initiated the development of a Decision Support Tool (DST) to assist land managers at the DoD Pōhakuloa Training Area (PTA) to spatially prioritize land management decisions associated with TER-S throughout a large and complex landscape.

The overall experimental design involved utilizing previously established management units where fences were constructed and nonnative feral ungulates removed at various time points in the past on the Island of Hawaii. In tropical montane wet forest, the research team utilized an existing ~20 year chronosequence of five feral pig removal units and adjacent feral pig present sites in Hawaii Volcanoes National Park and the Puʻu Makaʻala Natural Area Reserve (Cole and Litton 2014), and constructed a time-zero site to look at early impacts of nonnative ungulate removal in the Kahauʻalea Natural Area Reserve. In tropical dry ecosystems, they worked in three widespread vegetation types in the Pōhakuloa Training Area – Metrosideros woodland, Dodonaea shrubland, and Eragrostis grassland – where the experimental design for each vegetation type consisted of a set of existing ungulate removal units established at different points in the past and adjacent areas with feral ungulates present. They conducted vegetation measurements in all four vegetation types and detailed ecological processes measurements in three of the vegetation types (excluding the Eragrostis grassland). They also established a greenhouse nutrient manipulation experiment to test how nutrient availability affects native vs. nonnative invasive plant growth and competition (the only Go/No Go Point in the project). The results of the greenhouse experiment were then used to inform the establishment of a field nutrient manipulation experiment in TWF and TDF where the response of vegetation and key ecological processes to nutrient availability was assessed. Finally, a DST was developed through a series of facilitated meetings with PTA land managers to generate a knowledge base and develop a logic model to enable spatially explicit prioritization of management activities in relation to TER-S.

The response of vegetation to feral ungulate removal varied somewhat across plant communities in TWF and TDF, but general patterns were evident. First, ungulate removal is a key first step in conserving and promoting native plant communities. Second, nonnative invasive plants also seem to benefit from ungulate removal, making their control an important aspect to consider when fencing and removing ungulates. Third, recovery of TER-S occurred with ungulate removal in all four ecosystem types studied, and recovery was highest where there was a population present at the time of fencing and ungulate removal. Finally, the rate of recovery varied substantially across the studied ecosystems, with the most rapid responses occurring in tropical wet forest and the slowest in the TDF ecosystems. TWF ecosystems have higher resource availability (e.g., precipitation) which likely explains the more rapid response to ungulate removal in this ecosystem type. In addition, the TDF sites were more heavily invaded by nonnative plants at the time of fencing and ungulate removal, which likely also resulted in a slower recovery.

The response of ecological processes to nonnative feral ungulate removal also varied somewhat across ecosystem types, but several common trends were evident. Overall, there was little impact of ungulate removal on ecosystem C storage and flux. The exception to this trend was increased belowground C cycling in TWF with nonnative feral pig removal, and an increase in fine detritus (i.e., fine fuels) in dry ecosystems following ungulate removal. Ungulate removal resulted in an increase in nitrogen mineralization (i.e., increased nitrogen cycling) in TWF and Dodonaea shrubland, but not in Metrosideros woodland. Similarly, inorganic nitrogen availability increased following nonnative ungulate removal in TWF and Dodonaea shrubland, but not in Metrosideros woodland. In general, feral ungulate removal led to increased nitrogen cycling and availability, and increased soil nitrogen availability persisted over time following ungulate removal.

Subsequent greenhouse and field experiments tested if this increase in nitrogen availability favors nonnative invasive plants over native plants. In the greenhouse experiment, the research team found that all but one native species showed very little response to increased nutrient availability. In contrast, all nonnative plants and one native shrub responded positively to nutrient availability by increasing growth and whole plant C gain with increasing nutrient levels. The field experiment had more variability and a weaker response to nutrient manipulation, but this was expected given that nutrient manipulations take longer to manifest under field conditions. Taken together with the ecological processes responses, the results strongly suggest that elevated soil nutrient supply following nonnative ungulate removal negatively affects native plants while benefiting nonnative invasive plants.

Three of five steps in the DST development process were carried out successfully under this project. Using facilitated sessions with active participation by PTA environmental staff, the research team identified the decision question (What are the priority areas for management at PTA in regards to TER-S?), and a suite of spatially explicit criteria to inform the answer to this question. They also identified and assembled the base-layer maps to guide the overall DST process. Remaining steps in the DST include developing a flexible user interface and creating a spatially explicit map of decision scores to guide final management prioritization.

While nonnative ungulates degrade native vegetation and ecosystems, little prior work has been conducted to assess the response of vegetation and key underlying ecological processes to the increasingly common land management strategy of fencing and ungulate removal. The research conducted under this project provides insights into the effects of nonnative ungulate removal on plant communities and ecological processes on lands managed by DoD across the Pacific. In addition, the research team initiated a decision support framework for informing where management activities should be prioritized on the PTA landscape. As such, these results should be of direct utility for land managers on DoD and other facilities across the Pacific Island Region.