Articles | Volume 10, issue 3
Geosci. Model Dev., 10, 1131–1156, 2017
Geosci. Model Dev., 10, 1131–1156, 2017

Development and technical paper 16 Mar 2017

Development and technical paper | 16 Mar 2017

A joint global carbon inversion system using both CO2 and 13CO2 atmospheric concentration data

Jing M. Chen1,2, Gang Mo2, and Feng Deng2 Jing M. Chen et al.
  • 1International Institute of Earth System Science, Nanjing University, 22 Hankou Road, Nanjing, Jiangsu, China
  • 2Department of Geography and Program in Planning, University of Toronto, Toronto, Ontario, Canada

Abstract. Observations of 13CO2 at 73 sites compiled in the GLOBALVIEW database are used for an additional constraint in a global atmospheric inversion of the surface CO2 flux using CO2 observations at 210 sites (62 collocated with 13CO2 sites) for the 2002–2004 period for 39 land regions and 11 ocean regions. This constraint is implemented using prior CO2 fluxes estimated with a terrestrial ecosystem model and an ocean model. These models simulate 13CO2 discrimination rates of terrestrial photosynthesis and ocean–atmosphere diffusion processes. In both models, the 13CO2 disequilibrium between fluxes to and from the atmosphere is considered due to the historical change in atmospheric 13CO2 concentration. This joint inversion system using both13CO2 and CO2 observations is effectively a double deconvolution system with consideration of the spatial variations of isotopic discrimination and disequilibrium. Compared to the CO2-only inversion, this 13CO2 constraint on the inversion considerably reduces the total land carbon sink from 3.40 ± 0.84 to 2.53 ± 0.93 Pg C year−1 but increases the total oceanic carbon sink from 1.48 ± 0.40 to 2.36 ± 0.49 Pg C year−1. This constraint also changes the spatial distribution of the carbon sink. The largest sink increase occurs in the Amazon, while the largest source increases are in southern Africa, and Asia, where CO2 data are sparse. Through a case study, in which the spatial distribution of the annual 13CO2 discrimination rate over land is ignored by treating it as a constant at the global average of −14. 1 ‰, the spatial distribution of the inverted CO2 flux over land was found to be significantly modified (up to 15 % for some regions). The uncertainties in our disequilibrium flux estimation are 8.0 and 12.7 Pg C year−1 ‰ for land and ocean, respectively. These uncertainties induced the unpredictability of 0.47 and 0.54 Pg C year−1 in the inverted CO2 fluxes for land and ocean, respectively. Our joint inversion system is therefore useful for improving the partitioning between ocean and land sinks and the spatial distribution of the inverted carbon flux.

Short summary
A joint inversion system is developed for estimating the carbon fluxes in 39 land and 11 ocean regions of the globe using both atmospheric CO2 and 13CO2 stable isotope data. In particular, a biospheric model is developed to model both CO2 and 13CO2 fluxes over land to constrain the inversion. Relative to CO2-only inversion, the joint inversion system improved the partition between land and ocean carbon fluxes and possibly the distribution of the fluxes among land regions as well.