UBC Theses and Dissertations
Local-scale assessment of regional conservation plans : strategies for the conservation of animals and plants in a tropical rainforest and surrounding mixed-farmland Contasti, Adrienne Louise
There is an urgent need to conserve biodiversity in human-modified landscapes throughout the tropics. Animal conservation has traditionally focused on single species, but it remains unclear whether these strategies will also protect other taxa that co-occur within the ecosystem. These uncertainties can also affect plant conservation if management interventions change plant-animal interactions. I identified steps to mitigate the effects of hunting, forest product extraction, and farming on rainforest animals and plants in Suaka Margasatwa Buton Utara (SMBU) and the surrounding mixed-farmland on Buton Island, Southeast Sulawesi, Indonesia. First, I used Bayesian Network inference to assess whether protecting anoa (Bubalus spp.) habitat might also benefit other animals by modelling species co-occurrences in relation to habitat and human activities. Next, I completed a pantropical Network Meta-Analysis (NMA) to identify where controlling the foraging of granivorous mammals might reduce mortality of management-sowed seeds in human-modified forests. Finally, I used the NMA to guide an experimental assessment of how seed predation might affect the regeneration of nine plant species in the reserve and mixed-farmland. Buton macaques (Macaca ochreata) did not co-occur with anoa and were the only species to avoid human-dominated areas. The government might consider concentrating patrols in easy-to-access areas to increase the distribution of macaques throughout SMBU and the mixed-farmland. The NMA identified that granivore control could help reduce seed mortality, but which seeds to protect depended on the type of human activity that modified the forest. At SMBU, granivore control was not required in the mixed-farmland because seed predation was very low for 78% of the studied plants. Low seed losses in the mixed-farmland suggested that forest regeneration might be enhanced by increasing macaque distribution and natural seed rain throughout those areas. My approach could be used to design projects that conserve animals and plants in human-modified ecosystems. Local co-occurrence analyses can identify species that remain vulnerable to humans under conservation projects focused on other species. In data deficient situations, wide-scale evidence synthesis using NMA can help guide decision making, such as when to use granivore control, in human-modified ecosystems.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International