Exploring model uncertainty of the Congo basin rainforest carbon cycle

The Congo basin forest plays a crucial role in the global carbon cycle, contributing to more than 10% of the global terrestrial sink. This carbon potential appears more stable compared to other tropical forests. Historically, despite its importance for global climate, the Congo basin forest has received much less scientific attention than other tropical forests. Notwithstanding, in recent years the body of data and knowledge has reached a critical level that allows studying the carbon cycle of the Congo basin forest under the present climate.

The main objective of this ongoing research is to quantify the uncertainty related to the carbon cycle in process based models, and to decompose the different aspects contributing to this uncertainty, This will ultimately improve our understanding of the Congo basin carbon cycle within the present climate. We present results from two dynamic vegetation models (ED2 and FATES), which represent structural and functional heterogeneity of forests, over the Congo basin forest. We decompose the uncertainty related to model structure, climate drivers and model parameters. We focus on the data-rich site of Yangambi, located in the central Congo basin in the Democratic Republic of Congo. Both models are initialized with forest inventory plot data and driven with meteorological drivers from GSWP reanalysis. We evaluate the modelled carbon cycle on seasonal and diurnal time-scales against recent measurements from the Congoflux eddy-covariance tower (2020-2024). Preliminary results indicate that both models tend to underestimate observed net ecosystem carbon exchange, especially during the daytime. These results can provide new process-based insights as well as inform on the importance of model structure differences in modelling the carbon cycle over the Congo basin.