Understanding orangutan habitat use and connectivity across human-modified landscapes

In this thesis I focus on the orangutan (Pongo spp), a charismatic mammal, that is critically endangered and thus emblematic of the biodiversity crisis in Southeast Asia. Although orangutans have received a large amount of conservation focus and funding in the past, the orangutan population status outside of key study areas in intact habitat remains poorly understood and overall their numbers continue to decline. This dearth of information in human-modified landscapes in particular, presents a challenge to the conservation of the species, as the majority of the orangutan range is found outside of strictly protected areas and land-use change continues to reduce and fragment these habitats.

Here, I employ both empirical and theoretical approaches to investigate orangutan populations in human-modified landscapes. First, I focus on the landscape scale and use orangutan nest data to estimate orangutan density in a multi-use human-modified landscape in Sabah, Malaysian Borneo. I then use these estimates to investigate orangutan abundance across a habitat disturbance gradient (from continuous logged forest to remnant forest in mature oil palm) and assess the relationship between abundance and environmental covariates. I show that orangutans are found across the landscape, although at substantially reduced densities in remnant forest patches in oil palm, corroborating that for orangutan abundance forest habitat quality appears more important than isolation of patches.

At the local scale I use these density estimates, as well as published and expert-elicited data on orangutan ecology, to adapt and parametrise an individual-based model for orangutan populations in the landscape. I generate four plausible landscape management scenarios founded on existing environmental policies, sustainable certification standards and recommendations from published literature. By applying the individual-based model to these four management scenarios I was able to show that policies and certification standards maximising forest cover in human-modified landscapes may promote viable orangutan populations and facilitate movement among habitat patches, provided mortality during dispersal is minimised.

An impediment to understanding orangutans in human-modified landscapes, is a lack of targeted orangutan observational data. Data integration is a rapidly expanding field in ecology, which makes use of multiple different types of data sets to improve precision. I investigate if observations from surveys for mammal biodiversity, specifically orangutan bycatch on camera trapping images, can be used to augment targeted orangutan nest surveys through integration to improve precision whilst making use of existing data. Although I demonstrate the potential for data integration to be applied at the local scale for orangutans, ecological processes driving detection in different datasets, such as nesting behaviour and preference for terrestrial or arboreal movement need to be fully considered if this approach is to be successful.

Orangutan population viability assessments have been undertaken in the past, but crucially, these only consider dynamics in time and are not spatially-explicit. They are also limited to averaged parameters for vital demographic measures (e.g. emigration and immigration rates) across the species and do not take into account habitat outside large continuous areas of forest. With this in mind, in my last chapter, I scale up the individual-based model implemented at state level to the whole of Borneo and couple it with high resolution forest maps, to investigate the role small fragments play in maintaining orangutan populations and movement across the landscape. Using estimated offtake of orangutans from hunting, retaliatory killings and rescues, I also investigate how removing individuals from the landscape affects population dynamics and movement. The results reveal the importance of small remnant forest fragments in facilitating individual orangutan movements, with high numbers of individuals dispersing out of their natal area when small fragments are present, thus demonstrating a greater ability to expand their range. However, even moderate levels of added mortality (>2%) lead to steep declines in populations, highlighting the disproportionate impact of removing orangutans (i.e. by killings or translocations on populations).

This thesis is an important step towards a better understanding of orangutan conservation in the Anthropocene, adding to the increasing recognition that orangutan research and conservation focus need to expand to also include human-modified landscapes and support efforts to promote peaceful co-existence between orangutans and humans. The evidence presented and our abilities to forecast the effects of potential conservation initiatives are vital to better inform orangutan management and ensure positive conservation outcomes for the species in the future.