| Summary: | To better understand the role of terrestrial ecosystems in the global carbon cycle and their feedbacks to the global climate system, the predictability of ecosystem models that are used for quantifying net carbon exchanges between the terrestrial biosphere and the atmosphere need to be improved. My first part of dissertation is about reducing the uncertainty in the Terrestrial Ecosystem Model (TEM). I demonstrate how to reduce the model uncertainty from: (1) model parameter space; (2) model structure; (3) and model scaling up method in three studies. To reduce the uncertainty in model parameters, I developed an adjoint data assimilation system for TEM model. By assimilating surface carbon flux observations, the model parameters could be well constrained and improved. Based on observed nitrogen dynamics, I improved the model structure in representing nitrogen uptake mechanism. I demonstrated that the new model was able to better simulate the carbon and nitrogen dynamics of tundra and boreal forest ecosystems. For the model scale up method, I challenged the traditional model extrapolation method. By using spatially explicit dataset of plant photosynthesis from MODIS satellite, I developed a spatially explicit scale up scheme for TEM model, which has been proved to be superior to the traditional method. Global terrestrial carbon cycle could also be inferred from atmospheric CO2 concentrations with the help of atmospheric transport and chemistry model. Thus my second part of this dissertation focuses on atmospheric CO 2 inversion. Multiple atmospheric CO2 concentration datasets including surface measurements and mid-troposphere satellite retrievals are assimilated into an atmospheric chemistry transport model (GEOS-Chem) to estimate the sub-continental scale terrestrial carbon budget. In this part, I combine biogeochemistry modeling and atmospheric inverse modeling to have a two-step framework to improve the quantification of global carbon budget in terms of its magnitude and spatial distribution. The CO2 inversion ...
|
|---|