The Hawaiian Islands are both the extinction and invasive species capitals of the world. The result has been Hawaiian ecosystems fundamentally changed in form; that is, ecosystems replete with a mix of novel and native species. Most native Hawaiian plant species are bird-dispersed, yet no native avian dispersers remain in most Hawaiian ecosystems. Thus, ecosystem functioning will only be maintained by the handful of invasive vertebrate dispersers that now reside on the islands, most of which are birds. In this context, research efforts must shift focus to non-native bird species (and potentially rats) and the potential for these species to maintain native plant communities under current and predicted environmental conditions. To successfully manage and preserve Hawaiian terrestrial ecosystems, it is necessary to identify and characterize non-native invasive species that are dispersers of desired plant species, determine their role in ecosystem function, and improve non-native plant management plans, while facilitating the recovery of native threatened, endangered, and at-risk plants.

The overarching objective of this project was to determine the effects of plant disperser traits, competition, predation, and landscape features on native and non-native plant dispersal and recruitment. This was accomplished through the integration of field-based data collection, field experiments, and ecological modeling describing and quantifying seed dispersal in novel Hawaiian communities. Essential outcomes of the research include determining how well different species of non-native birds disperse native plant species across environments, whether non-native rat species have a cumulative positive (via seed dispersal) or negative (via predation on bird seed-dispersers and/or via seed predation) impact on communities, and creating predictive models to be used for management in novel environments and under future abiotic and biotic scenarios. This research will provide Department of Defense (DoD) and the scientific community with the essential tools for managing and maintaining native plant communities in Hawaii and other Pacific Islands.

The overall technical objective was to quantify and predict how novel communities of seed dispersers structure plant communities across environmental gradients. To address this overall objective, the project team:

• Determined how environmental variables affect networks of non-native seed dispersers. The project team accomplished this by documenting variation in seed dispersal network structure (number and type of plant and disperser species) across ecological contexts, estimated the effects of abiotic factors and species interactions on the persistence of disperser species, and determined the impacts of competition on seed preference and dispersal.
• Determined how the movement of non-native dispersers, the availability of fruits, and seed dispersal affected plant communities. The project team accomplished this by assessing how traits of seed dispersers influenced seed viability and seedling recruitment and constructed agent-based movement ecology models to examine how disperser behavior and landscape features affect seed movement and plant community structure.
• Predicted the effect of changing environmental conditions on the networks of non-native seed dispersers and plant communities. The project team constructed predictive ecosystem functioning models using output from Objectives #1 and #2. This extended the benefits of the research beyond the study sites on Oahu to inform future management action in new landscapes and under future scenarios of biotic and abiotic (e.g., climate change) shifts in other locations in Hawaii and the Pacific. Examples include the potential for new dispersal events given the landscape composition, risk of invasion, and changes to functioning with the addition or deletion of dispersers.

Multi-species interactions are crucial to the maintenance of ecosystem structure and function. This is especially true for seed dispersal networks, where interactions among seed dispersers, plants, and predators influence the efficacy of seed dispersal and ultimately, community structuring. Research on seed dispersal by birds has largely been restricted to individual species and focused on the identification of seeds in their diets and their role in plant dispersal. Thus, most current research does not reflect the reality of multiple interacting native and non-native species (birds and rats), relative disperser effectiveness, and the capability of non-native dispersers to maintain native plant communities in severely altered ecosystems. This project incorporated these interactions through the comprehensive examination of seed dispersal networks comprised of seven non-native invasive bird species, two common non-native invasive rat species, seven non-native invasive plant species, and eight ecologically important native plant species. Collection of field-based variables included species abundance, reproduction, predation, inter- and intra-specific competition, diet preferences, disperser behavior, gut passage times, and seed germination rates to create and parameterize movement ecology models for assessing seed dispersal across heterogeneous landscapes. When integrated with high resolution vegetation and elevation data in a landscape modeling framework, these individually-based movement ecology models can be used to estimate and predict how interactions within seed dispersal networks influence dispersal and recruitment probability of native and non-native plants.

This project examined multiple seed dispersal networks comprised of interacting native and nonnative plant and vertebrate species across several trophic levels and ecological contexts in order to develop predictive models for assessing recovery and maintenance of key ecological processes including dispersal, recruitment, and establishment of threatened, endangered, and at-risk plant species. These predictive models, capturing a range of elevation and precipitation gradients and utilizing generic bird (e.g. mass, life history, dietary) and plant (e.g. size, color, fruiting height) traits, are particularly useful as they can be applied to other similar Pacific Island habitats with different bird and plant species. Importantly, this work will provide DoD and adjacent resource managers with the tools to go beyond current non-native invasive species control and eradication efforts in Hawaii by providing meaningful predictive measures of native and non-native plant dispersal and establishment. In this context, the results will put DoD in a powerful position to match environmental objectives with agency needs, and to provide leadership on responses to global climate change and the recovery of native threatened, endangered, and at-risk plants in the Pacific Islands.