Seminar: Hoontaek Lee

Interactions between the carbon and water cycles are a key aspect of understanding and projecting the global carbon cycle and climate. However, current land model simulations are less constrained in this regard, especially at the interannual time scale. Incorporating observational data into land models is a promising way to reduce uncertainties, such as those arising from parameters and model structure, and to advance our understanding of carbon-water interactions. In this talk, I will summarize two studies using a model-data integration framework, SINDBAD.

The first study diagnoses regions that dominate the variance in globally integrated terrestrial water storage (TWS) interannual variability (IAV). Error hotspots are primarily identified in tropical regions. Further analysis indicates that these hotspots are associated with lateral flow dynamics. These findings suggest regions and processes that could lead to the efficient model development. In the second study, a series of land carbon cycle processes are integrated into the SINDBAD model. By using the model in conjunction with observational constraints and an ensemble of land model simulations, the variability in global net ecosystem exchange and gross primary productivity is attributed to regional variability. The findings highlight tropical humid regions as key areas for the IAV of global land-atmosphere CO2 exchanges and the inter-model spread in its simulations.