Evaluating the CarbonTracker 2015 Posterior Suite Against Multiyear North American Observed Biogenic Fluxes

Atmospheric carbon dioxide (CO₂) significantly influences the radiation budget of Earth and therefore contributes to the greenhouse gas climate forcing. Global atmospheric mixing ratios of carbon dioxide have increased by 36% in the last 250 years, with half of this rise occurring in the last four decades.  Levels of atmospheric CO₂ are moderated by terrestrial and oceanic uptake, with approximately 55% of the annual anthropogenic emissions released globally sequestered by the land and ocean. Because the North American terrestrial biosphere acts as a large sink for atmospheric CO₂, it is important to determine its spatial distribution and quantify of its magnitude. Accurate and precise quantification of these terrestrial fluxes is critical for projecting future atmospheric concentrations of CO₂.

Extensive evaluations of carbon fluxes show considerable variation in terrestrial uptake of atmospheric CO₂ year-to-year. Interannual variability of atmospheric CO₂is strongly driven by both climatic processes and the terrestrial carbon cycle. Variability in the magnitude of carbon exchange in space and time is governed by meteorology and ecology, as well as the length and timing of the growing season. There exists a large spread of projected future warming in coupled carbon cycle-climate simulations, and carbon respiration and uptake processes in terrestrial biospheres are likely to be affected as a result of climate change. This underscores the importance of improving understanding of land-atmosphere carbon exchange.

Atmospheric inversions are a primary method for estimating carbon fluxes on a global or regional scale. Many current inverse estimates use as their priors the surface flux posteriors provided by the CarbonTracker NOAA data assimilation system. The product made available in 2015 (hereafter referred to as CT2015) uses atmospheric CO₂ mixing ratios from the NOAA global network of stations to optimize carbon fluxes at the surface over large ecological regions. These fluxes are computed at a 1°x1°horizontal spatial resolution and a 3 hourly time step over North America. The CT2015 provides fluxes from 2000 – 2014.

CT2015 is derived as a mean of eight different posterior estimates. This suite is developed using two different terrestrial priors (land biosphere from GFEDv4.1s; GFED CMS), ocean priors (Takahashi pCO₂ climatology; fluxes from the ocean inversions project), and fossil fuel priors (emissions from Miller/CDIAC; ODIAC). Here, we evaluate the suite of optimized terrestrial biogenic fluxes over North America from 2000 – 2014. To date, the range and variability of these eight posteriors over space and time have received little evaluation against observations. The goal of the present study is to determine how CT2015 biogenic fluxes change with space and time, and how representative they are of measured fluxes. We show how the range of CT2015 posteriors compares to model mean-data differences, and how this comparison varies throughout North American sub-regions and over the 15 years examined.