A global understanding of forest-savanna transitions

Savannas dominate the tropics, and often occur in areas which could be forest, given the rainfall and soils. Why savannas are where they are, and how that might change under climate change has long preoccupied scientists, but a general understanding has yet to be developed. Remote sensing offers a way to address this problem, and new data sets from lidar and radar sensors in particular, have the potential to describe and explain savanna-forest transitions globally.

Given the importance of these ecosystems for local people, carbon cycling and biodiversity, this understanding is vital for responding to the climate and biodiversity crises. In this PhD you will use these new remote sensing data and machine learning to understand how and why savannas transition into forests and how this is changing. You will map and classify the world's savanna-forest transitions and test what determines their location and temporal stability.

Theoretical models suggest that both gradual and abrupt transitions between savannas and forest can occur, depending on the traits of trees. However, in the real world, it is not clear why in some places, gradual transitions occur, whilst in other, there is a mosaic of sparse savanna and closed forest. Nor do we understand how stable these patterns will be under climate change, changes to herbivory (such as the loss of elephants) and CO2 fertilisation.

Filling these gaps in our knowledge is important because savannas are thought to be particularly sensitive to climate change: they are expected to both be a globally important carbon sink, and to change rapidly as CO2 fertilisation allows trees to outcompete grasses. Which factors control savanna[1]forest boundaries, and how these factors will respond to climate change and other environmental changes, will therefore be vital for the future of savannas.