Tropical forests and savannas are ecosystems of great ecological and economic importance. Predicting their response to future climate change is therefore crucial, but represents a challenge because the ecological functioning that determines their distribution remains controversial. One theory suggests that hydroclimate (rainfall, relative humidity) is the main factor determining the distribution and transitions between forests and savannas. Another theory suggests that these ecosystems represent alternative stable states, meaning that they coexist under the same climatic conditions and are stabilised by feedback loops between fire and vegetation. Fire therefore plays a central role in their maintenance or transition from one state to another. However, long-term empirical evidence testing the role of fire on these ecosystem transitions remains limited. Resolving this debate is essential for anticipating the future dynamics of tropical landscapes and guiding conservation and management strategies.

The aim of this project is to test whether changes in fire regimes alone can trigger transitions between forests and savannahs in West and Central Africa, or whether other factors (climate, humans) are decisive. If fire is the main driver, these transitions would not be directly reversible, which would imply that fire management could be essential to landscape stability and restoration. Since these transitions are slow processes, they require long-term observations on time scales of more than a thousand years, which only palaeoecology can provide. However, palaeoecological reconstructions present methodological challenges.  

To address this issue, we propose to combine the FTIR analysis of charcoal with the calibration of a fire model, in order to quantitatively reconstruct the main components of past fire regimes (intensity, frequency, area burnt, origin). These analyses will be carried out on sediments covering the last 3,000 years, and will be compared with reconstructions of vegetation and climate in order to establish the precise chronology of their changes.

Finally, a dynamic tree-cover model will be developed using these reconstructions to assess whether changes in fire regimes alone can trigger forest-savannah transitions or whether other factors need to be taken into account. These simulations will also be used to determine whether alternative stable states exist in relation to tree cover and whether the changes in vegetation observed correspond to tipping points.

By combining quantitative fire reconstructions with dynamic vegetation modelling, this project will establish a new methodological framework for studying past transitions between ecosystems and identifying their drivers. Our results can be used to improve predictions of the responses of tropical ecosystems to global change, and will ultimately help to guide future management and conservation strategies for these ecosystems.