My research uses the information of these boundary conditions both offline or in an interactive manner to explore the extent to which climate can be predicted. For instance, in a recent paper published in the journal Science , I used a coupled atmosphere-land-vegetation model to study the interannual and interdecadal climate variability in the Sahel where it had undergone a long drought from late 1960s to the 1980s. This work demonstrated the importance of a dynamic vegetation and soil moisture in enhancing this multi-decadal variation initiated by a change in the global SST pattern. I am also concerned about human impact on the climate and ecosystem such as deforestation and desertification. I believe these effects can not be well understood without a deep understanding of the natural variability at first place. I thus try to study both natural and anthropogenic changes. For this purpose, an extensive use of models and comparison with observational data have become essential.
Many mysteries remain in the global carbon cycle, such as what caused the cycles in the glacial-interglacial CO2 variation, and the `missing carbon sink' which is yet to be accounted for in order to close the present-day carbon budget. My current research aims to contribute to our understanding of these questions. A less emphasized issue is the role of terrestrial biosphere which exhibits high heterogeneity in space from individual plant to continental scales, and multiple time scales from minutes for leaf-level transpiration to many thousands of years for soil carbon accumulation. I have developed a model called VEGAS (Vegetation Global Atmosphere and Soil) to address vegetation dynamics and biospheric carbon cycle. This is has been coupled to the HAMburg Ocean Carbon Cycle model (HAMOCC) to address both the glacial-interglacial and present-day carbon-climate interaction.
Models developed in house include:
VEGAS (VEgetation Global Atmosphere and Soil): vegetation dynamics and terrestrial carbon cycle; coupled to SLand.
SLand (Simple-Land): Land surface energy and water budget, flux exchange with lower atmosphere.
QTCM (Quasi-equilibrium Tropical Circulation Model): a simple GCM, developed in collaboration with D. Neelin at UCLA; in collaboration with R. Iacono (ENEA/Italy), a global version of this model has been developed.
In addition, these have been coupled to the HAMburg Ocean Carbon Cycle model (HAMOCC) (developed at the Max-Planck Inst. for Meteorology) and a simple physical ocean model.
We have an Earth system model that simulates the important geophysical and biogeochemical processes, yet fast to run and easy to analyze. In the past, scientists have been somewhat limited to either highly complex models such as GCMs, or extremely simple models such as energy balance models (EBMs). Thanks to the improved understanding of the climate system, and to the fast increase in computing power, it has become possible to have a reasonably sophisticated model with balanced complexity among the individual components, yet much faster than GCMs (For instance, QTCM is about 200 times faster than CCM3 at similar resolution; yet it performs well in simulating climate variations such as El Nino influence on global precipitation.). Indeed, most of the models described above have been developed/run on a laptop computer runing linux. In addition, we also use semi-empirical models as well as GCMs, depending on the problem in hand. These models are enabling us to tackle some long-standing issues in paleoclimatology and climate change.