Savannas play a crucial role in providing essential ecosystem services, including carbon sequestration, biodiversity conservation, and water regulation. However, these ecosystems face threats from changing climatic conditions and unsustainable rangeland practices. Intensive livestock grazing has led to widespread degradation of savanna systems, characterised by shrub encroachment and loss of ecosystem functions. Shifting rangeland use from cattle grazing to native herbivores is proposed to reverse these negative trends, but the consequences for ecosystem properties and processes are not yet clear. Climate change and extreme events such as droughts exacerbate the challenges faced by savannas, potentially affecting the provision of essential ecosystem services on which local people’s livelihoods depend. Ensuring savanna resilience to environmental stressors, especially increased aridity, is essential for sustaining these vital services. This resilience largely depends on the interaction between herbivores, vegetation composition, and climate. Unfortunately, our understanding of this interaction is limited, and current research has not addressed how savannas will respond to uncertain future climate conditions under different rangeland management practices. In this thesis, I used the ecohydrological simulation model EcoHyD to investigate the effects of different rangeland practices on vegetation and water resources of a semi-arid savanna under current and future climate. This research makes two important contributions, first by examining the impacts of different future climate projections at a regional scale, which has not been the case in previous studies, and second by modelling rangeland types beyond grazing. In Chapter II of the study, I examined the impacts of grazing and browsing herbivores on plant diversity and ecosystem functioning. The results showed that intensive grazing negatively affects grass cover and water availability, whereas browsing herbivores increased vegetation cover, plant functional diversity, and water use. However, these outcomes may depend on herbivore community composition and climate contexts, especially when extreme events Summary like prolonged droughts occur. In Chapter III, I addressed these context-dependencies by investigating how different rangeland management options (i.e., grazer-dominated, mixed- feeders or browser-dominated herbivore communities) impact savanna ecosystems during droughts of varying durations. I found that diverse herbivore communities with a higher proportion of browsers combined with high plant functional diversity improve plant community resistance to drought and recovery after drought events, thus leading to enhanced ecosystem functioning and resilience. However, to fully understand ecosystem responses to climate change, temperature increases, precipitation changes and droughts need to be considered together. In Chapter IV, I therefore analysed the potential use of wild herbivore communities in regional climate adaptation plans to ensure the long-term resilience of savanna rangelands to the impacts of all aspects of uncertain climate change. My results have shown that while climate change alone does not necessarily lead to ecosystem degradation, in combination with poor management practices such as intensive grazing it can quickly lead to critical thresholds being exceeded. On the other hand, I have found that reducing the density of grazers and including mixed herbivores and browsers can promote ecosystem stability and resilience. Crossing tipping points can thus be delayed, or even avoided. In conclusion, my research suggests that utilising diverse herbivores and functionally diverse plant communities in rangeland management strategies can improve ecosystem resilience, minimise the risk of irreversible degradation, and better manage uncertainties associated with climate change.
