Landscape dynamics in the Miombo woodlands of Southeast Angola

Land cover changes, driven largely by anthropogenic activity, profoundly influence socio-ecological landscape dynamics. These alterations can reshape the structure and function of ecosystems, affecting their capacity to support biodiversity, regulate climate, and provide essential services. Human-induced modifications to the Earth's land surface not only modify ecological processes but also have wide-ranging consequences for human well-being, highlighting the interconnected nature of environmental and societal systems. The highlands of southeast Angola host a vast expanse of sparsely populated miombo woodland that is globally significant as a carbon store and for biodiversity conservation. These woodlands also hold immense hydrological importance for southern Africa and serve as a vital resource for local livelihoods in Angola, providing essential provisioning ecosystem services. However, anthropogenic pressures may be impacting the functioning of the ecosystems in this region. The miombo woodlands and grasslands here offer a critical case study for understanding the interplay between natural processes and anthropogenic influences on tropical grassland, savanna, and woodland ecosystems due to the heterogeneity in vegetation types and land use intensities. This thesis investigates land cover dynamics, fire regimes and impacts within wooded ecosystems, and changes in above-ground biomass to inform sustainable management and biodiversity conservation in the southeast of Angola, so that the miombo woodlands in this region can continue to provide a wide range of ecosystem services. Analysis of land cover dynamics over space and time can identify where conversion or modification of land cover is occurring and provide insights into possible associations with known causes and drivers of change. In chapter 3, we developed a 30-year land cover time series using data from the Landsat archive. We found that total woodland extent remained roughly stable despite fluctuations between dense and open woodland structures. Furthermore, we observed a peak in canopy opening towards the end of the Angolan civil war which we associate with the resettlement of displaced people. Over 30 years, 61% ± 2% of canopy opening was offset by subsequent canopy closure, which peaked a decade after the war ended, indicating the resilience of miombo systems. A resource-use frontier in the northwest highlights urban-driven pressures on woodland resources, while remote core areas of woodland remained relatively stable. These findings can support conservation managers and policymakers in spatially locating potential area-based forms of conservation in support of the 30 x 30 target of the Global Biodiversity Framework and support the sustainable management of the woodlands in the face of woody resource use frontiers driven by urban demand. Fire, though a pivotal component in the functioning of African mesic savannas, is often maligned as destructive in tropical wooded ecosystems and is attributed as a cause of woody vegetation losses. Despite this, some authors argue that fire does not cause the loss of woody vegetation and is instead a consequence of vegetation structure. In chapter 4, using burned area data from MODIS, the land cover time series from chapter 3 and methods for causal inference, we found that fire is both a cause of canopy opening and a consequence of prior canopy opening, creating a feedback loop where fire can amplify itself within the woodlands, with significant implications for carbon dynamics, biodiversity, and local livelihoods. Late dry-season fires, in particular, exacerbate canopy loss, underscoring the need for sustainable land use practices to mitigate runaway land cover change. These findings contribute to the ongoing debate surrounding the impact of fire in tropical savannas, but will also help to inform fire policy and management in southeast Angola to reduce the impact of self-amplifying fires that can cause unintended land cover changes. Woody encroachment threatens biodiversity and ecosystem function across tropical savannas, including Africa’s largest wooded savanna, the miombo woodlands, where grasslands are rapidly declining. While woody encroachment and densification increase carbon sequestration, a priority under the Paris Agreement, they come at the expense of open-ecosystem biodiversity and services. In chapter 5, using ALOS ScanSAR radar data and level 4 biomass estimates from GEDI, we estimated above-ground biomass from 2018 – 2023 and identified woody encroachment onto grasslands and densification within woodlands in southeast Angola. We found that both processes are widespread and progressing rapidly, with woody encroachment increasing at 2% annually as a proportion of grassland and densification occurring at 3% annually as a proportion of woodland. Densification contributed 21 times more biomass gains overall and twice the per-hectare biomass increase compared to woody encroachment. Both processes are linked to remoteness from anthropogenic land use, climate change, and fire frequency, with woody encroachment more common at lower fire frequencies and where late dry-season fires occur. We argue that prioritizing the conservation of grasslands over carbon gains is essential to preserving biodiversity, while carbon-focused management objectives should target the woodlands for greater efficiency. Collaborating with local communities to align conservation, livelihoods, and fire management may help limit encroachment and protect the region’s grasslands.