| dc.description.abstracteng | Insect pollination is an economically important ecosystem service that depends heavily on wild pollinators. Landscape transformation caused by conversion to agriculture threatens habitats of wild pollinators, reducing their potential to provide ecosystem services. The landscape in Jambi Province, Sumatra, Indonesia, is an example of a region undergoing landscape transformation, from biodiverse natural forests, to intermediate land uses such as fallow shrubland and jungle rubber, to monospecific oil palm plantations. My dissertation explores how transitions between these land uses impact pollinator biodiversity and ecosystem functions and services. I consider different facets of this interface: the impacts of adjacent forest on pollination services to oil palm; mechanisms driving ecosystem functions and services following agroforestry enrichment within oil palm; and landscape-scale comparisons quantifying interacting local and landscape effects on a native pollinator species.
My first chapter reviews the state of the knowledge of oil palm pollination by insects. I conducted a systematic review of biotic and abiotic drivers of oil palm pollination and pollinator populations. I present the current understanding of the globally introduced West African Weevil (Elaeidobius kamerunicus), whose regional population fluctuations have negatively impacted yield and resilience, as well as other potential pollinator species endemic to particular growing regions. Based on my review, I describe specific issues concerning biotic, management, and climate drivers of pollination that should be the focus of future oil palm pollination research.
In my second chapter, I examine the role that natural habitat can play as a source of pollination ecosystem services and ecological spillover effects in an oil palm field experiment. I compared treatments controlling visitors and pollination of female oil palm inflorescences over a distance gradient from an adjacent forest border. I found that exclusion of all visitors significantly decreased fruit set compared to other treatments, confirming insect pollination is necessary for adequate yield. Forest proximity had a significant positive effect on fruit set when large visitors were excluded. This effect was not significant for treatments that minimized pollinator contributions, suggesting this trend was not due to abiotic factors. However, insect abundances associated with oil palm inflorescences were not strongly related to distance from forest, and only E. kamerunicus had a significant relationship with fruit set. These results could suggest that non-consumptive ecological spillover from forests may influence oil palm pollination, though more experimental work is needed to identify these interactions.
In my third chapter, I explore how enriching the oil palm agricultural matrix with up to six tree species played a role in driving insect-mediated ecosystem functions. Within a plantation-scale, long-term oil palm biodiversity enrichment project, I disentangled the direct and indirect effects of enrichment on services (pollination, biocontrol) and disservices (herbivory) using random partition analysis and structural equation models. These models indicate that changes in canopy openness, driven by enrichment treatments, played an important role in ecological patterns at multiple levels of interaction. These had effects on herbivorous insects and pollinators, the latter of which had a positive effect on the fruit production of phytometer plants (Capsicum annuum) placed in the plots. Our results show that, even in early stages of ecological restoration of oil palm, ecosystem functions and services can be affected. These effects are mediated by decreasing canopy openness; however, these relationships may change as enrichment communities continue to develop.
In my fourth chapter, I examine the counterbalancing roles of local land use and amount of landscape habitat in pollinator survival and growth, using the native stingless bee Tetragonula laeviceps as a focal species. I established three colonies in 40 plots within Jambi’s transformation landscape, selecting from four predominant habitat types (degraded primary forest, shrubland, rubber plantation, and oil palm plantation) and controlling for a gradient of natural habitat (i.e., forest and shrubland) composition in a 500 m radius. I found that hives with higher local flower species richness had more pollen stores, which in turn was associated with increased bee and colony size. Colonies in structurally complex habitats such as forest and rubber plantations had lower mortality and greater gains in hive size than colonies in open habitats such as oil palm and shrubland; however, open habitats had higher flower species richness and abundance. Open habitats, which are increasing with rainforest conversion, reduce suitable nesting habitat but may increase floral resources in the landscape. Considering the key pollinating function of stingless bees, understanding how this trade-off translates to landscape and population scales will be critical in light of the continued deforestation crisis in the tropics.
In summary, understanding and protecting the resources supporting pollination are critical to improving the sustainability of oil palm and can support the livelihoods of people living in the landscape. The impact of habitat loss from agriculture conversion can adversely impact wild pollinator populations through the loss of suitable habitats, as I observed with T. laeviceps, but also changes the flowering resource landscape. In my review of oil palm pollination, I show that a better understanding of biodiversity and landscape drivers of pollination of oil palm itself is needed. My own oil palm field experiment demonstrates how natural habitat and biodiversity may interact positively with oil palm production. I show in my experiments within the oil palm biodiversity enrichment project that tree biodiversity enrichment can indirectly influence pollination services and ecosystem functions, even at an early stage. However, the ecosystem services and disservices affected by ecological restoration will continue to change as the ecological community develops. Our task in future research will be to continue to disentangle these relationships with the aim of recovering or preserving biodiversity and ecosystem function while informing sustainable ecological strategies for farmers and land managers. | de |