Seminar: Aleksei Lipavskii

Drylands occupy up to 40% of the land surface and contribute significantly to the long-term trend and interannual variability of the global carbon cycle. However, accurately simulating ecosystem carbon and water fluxes with global dynamic vegetation models remains a significant challenge. Here, we evaluate the QUINCY model against observations from a selection of FLUXNET sites in drylands and highlight limitations in its ability to simulate dryland carbon and water fluxes. We then attempt to reduce the model-data mismatch. Firstly, we reduce the mismatch in simulated phenology by forcing the model with remotely-sensed leaf area index. Secondly, we address the remaining biases in the water cycle simulation by identifying missing soil hydrological processes (such as hydraulic redistribution and ponding), which are important for reproducing seasonal soil moisture dynamics. We assess the improved model performance using observed fluxes and derived ecosystem characteristics (e.g. water-use efficiency and evaporative fraction metrics) and test the robustness of the results using a parameter sensitivity analysis.