Social-ecological dynamics at the livestock-wildlife interface in an African savanna

Amidst the global biodiversity crisis, eastern and southern Africa's rangelands stand out as one of the most critical regions for conservation attention. They contain much of the world's biodiversity and some of the only intact large mammal assemblages. However, over the next 50 years, these ecosystems will be disproportionally threatened by the global-scale impacts of land transformation and climate change. So far, the primary response to the global biodiversity crisis, both globally and within eastern and southern Africa, has been through an increase in the extent of protected areas (PAs). Yet, PA expansion has been unable to halt biodiversity loss. This thesis explores the factors driving this negative ecological trajectory at the livestock-wildlife interface in Africa's rangelands. First, I use a Kenyan case study to demonstrate that the current PA network inadequately represents biodiversity, indicating that conservation practice and policy need to extend across "working landscapes", such as rangelands and agricultural areas, to prevent loss. Second, I create a metric to show, for the first time, that PAs in eastern and southern Africa are under immense pressure along their boundaries, threatening their utility. Because of this, I argue that we must develop conservation strategies that cut across species and land-use boundaries to alleviate pressure on PAs and ensure that global biodiversity targets are met. To understand how best to alleviate these pressures, I use a social-ecological systems (SES) lens to identify the drivers of biodiversity loss at the livestock-wildlife interface, using data collected across six ecosystems in southern Kenya. To do so, I compiled a wide variety of socio-economic and ecological data. This includes land price data collected from scraping websites, data from multiple sources of long-term changes in climate, rangeland productivity, and wildlife and livestock numbers, and the collation of data on ecosystem boundaries and land tenure. Specifically, I collected and synthesised one of the most extensive data sets on the status of any rangeland by mapping the extent of livestock enclosures, fencing, and agriculture across 35,000 km2 in southern Kenya. My analysis of these data demonstrates how the profit-driven turnover of ownership, subdivision and transformation of land in southern Kenya is rapidly underway, fragmenting ecosystems, changing governance structures, and degrading rangelands. This process is causing widespread detrimental impacts to savanna biodiversity and the sustainability of the pastoral social-ecological system. Our comprehensive analysis exposes proximate anthropogenic stressors as overwhelmingly strong drivers of ecosystem-level change, currently eclipsing the impact of climate change, and confirms their importance for conservation within the rangelands of eastern and southern Africa. The lesson from southern Kenya is that African savannas can quickly flip from rain-driven to human-driven ecosystems. Thus, conservation solutions must focus on local drivers of change, which must be traced back to the higher-level socio-economic and political developments to which they are responding. In the face of global change, it is evident that economic incentives, underwritten by devolved management and governance of rangelands and wildlife, will be vital for conservation in this region. Ensuring there is an instrumental value of wildlife for landowners and national governments is critical, allowing landholders to convert wildlife and biodiversity from a short-term liability into a long-term asset. A landscape-based approach working across scales and sectors while incorporating and reacting to these SESs complexities is a critical step in addressing these shortcomings. I provide an example of this approach in southern Kenya, in what we term "inside-out" conservation. This broader perspective improves the chances that global biodiversity targets and sustainable development goals are aligned. It also ensures that limited conservation resources are used to benefit both people and wildlife through PAs and in the larger, more functional landscapes in which they are embedded.