Estimations of global terrestrial carbon flux using bottom-up and top-down approach

Estimating global terrestrial carbon fluxes with high accuracy is important to understand global environmental changes. Moreover the estimations of the global spatiotemporal distribution may contribute to the political and social activities. In order to reveal the present state of terrestrial carbon fluxes covering a wide and a decadal scale, using the satellite-based diagnostic biosphere model is suitable because of uniformly observing on the present global land surface condition. However the satellite-based diagnostic model can simulate the terrestrial carbon flux in only a few decadal periods and cannot calculate back to 100 years ago. Therefore the model estimations have the potential to underestimate the annual terrestrial carbon fluxes as a result of doing spin-up to the steady state. Because the steady state is a pre-industrial era, not present day, so the flux becomes to zero throughout the global scale. In this study, we optimized the spin-up time of the terrestrial biosphere model (BEAMS) in each sub continental region using estimations of carbon fluxes by the atmospheric transport model (GOSAT L4A global CO2 flux). First, we made the BEAMS steady state by the spin-up run for 10000 years, and then the carbon pools in the biomass and soil was adjusted to fit the GOSAT carbon flux. Significant improvement of the estimation accuracy was achieved by using the two satellite observation data (GOSAT as atmospheric information, and MODIS as land surface information). We evaluated our new carbon flux estimations on various spatial scales. Annual global carbon fluxes were indicated similar values between BEAMS, GOSAT L4A, and GCP estimations, and perhaps these may be reasonable. In a tropical regions that are low satellite observation data, the temporal patterns of the carbon flux was indicated various changes, so the accuracy of carbon fluxes in some regions remained a matter of discussion.