This is not the latest CarbonTracker update! Link to latest.
Executive Summary (CT2008)
We present the third release of a combined measurement and modeling system that keeps track of the emissions ("sources") and removal ("sinks") of atmospheric CO2 globally from January 2000 through December 2007. CarbonTracker 2008 (released on 31 October 2008) incorporates several improvements over the previous release.
  1. We have added a number of observation sites. In collaboration with the Commonwealth Scientific and Industrial Research Organization (CSIRO), Australia, measurements of flask samples collected at Cape Grim, Tasmania, Cape Ferguson, Queensland, Macquarie Island in the Southern Ocean, Mawson, Antarctica, and Casey, Antarctica, were added. In North America continuous measurements at Southern Great Plains, Oklahoma, with Lawrence Berkeley National Laboratory (LBNL), and at Egbert, Ontario, with Environment Canada (EC), were added
  2. All CO2 observations used in CarbonTracker to estimate sources/sinks are now available for download in the form they were used.
  3. Optimized global time-dependent fields of the CO2 mole fraction are now available for download at three hourly resolution.

Figure 1: The long term mean biological uptake

Figure 2: The long term mean fossil fuel emissons

Estimates of CO2 sources and sinks
From 2000 through 2007 ecosystems in North America have been a net sink of -0.72 ± 0.59 PgC/yr (1 Petagram Carbon equals 1015 gC, or 1 billion metric ton C, or 3.67 billion metric ton CO2), offsetting about one third of the emissions of 1.98 PgC/yr from the burning of fossil fuels in the U.S.A., Canada and Mexico combined. Our estimates include sub-continental patterns of sources/sinks coupled to the distribution of dominant ecosystem types across the continent (see Figures 1 and 2). The sinks are mainly located in the agricultural regions of the Midwest (36%), deciduous forests along the East Coast (33%), and boreal coniferous forests (17%). There also appears to be substantial year to year variation of the sink, with a range of -0.1 to -1.0 PgC/yr, correlated with large scale temperature and moisture variations. We estimated the lowest net annual uptake by ecosystems in 2002, when there were widespread drought conditions through most of the West. In contrast, our observing system did not detect an effect from the 2007 drought in the Southeast. This is likely due to lack of coverage of the area (Figure 3) in our current observing network. The estimates are optimally consistent with measurements of ~18,000 weekly air samples from across the world, ~20,000 daily averages of continuously measured CO2 at sites located primarily in North America and ~9,000 daily averages from tall towers at three locations within the continent (see Figures 3 and 4).

Figure 3: North American sites only
Word of caution about the biological flux maps
Figure 1 shows 1 x 1 degree detail for estimated fluxes. With the present observing network of about a dozen sites the detailed 1 x 1 degree fluxes should not be interpreted as quantitatively meaningful for each block. To spread the influence of sparse observing sites we make the assumption that large ecosystem regions respond in the same way to variations of temperature and light. However, temperature and light are not uniform in an entire region, and thus the same response function does not produce a uniform flux over the region. Thus we caution that the spatial detail is only predicted by CarbonTracker based on the assumption of large-scale ecosystem coherence, but has not been verified by observations.

Calculated time-dependent CO2 fields throughout the atmosphere
A "byproduct" of the data assimilation system, once sources and sinks have been estimated, is that the mole fraction of CO2 is calculated everywhere in the model domain and over the entire 2000-2007 time period, based on the optimized source/sink estimates. As a check on model transport properties, calculated CO2 mole fractions were compared with measurements of ~23,000 air samples taken by NOAA/ESRL at 30 aircraft sites, which had not been used in the estimation of sources/sinks. Column averages of the CO2 mole fraction have been calculated as well, and they can be compared to satellite measurements of the same quantity when the averaging is done in the same way as for the satellite results.

Figure 4: All observation sites used in CarbonTracker
It is important to note that at this time the uncertainty estimates for the sources/sinks are themselves quite uncertain. They have been derived from the mathematics of the data assimilation system, which required several "educated guesses" for initial uncertainty estimates. The paper describing CarbonTracker (Peters et al. (2007), Proc. Nat. Acad. Sci. vol. 104, p. 18925-18930) presents different uncertainty estimates, based on the sensitivity of the results to 14 alternative yet plausible ways to construct the CarbonTracker system. For example, the 14 realizations produce a range of estimates for net annual mean terrestrial uptake in North America from -0.40 to -1.01 PgC/yr, as given in the PNAS paper. The procedure is described in the Supporting Information Appendix to that paper, which is freely downloadable from the PNAS web site. In addition, the estimates do not take into account several additional factors noted below. The calculation was set up for sources/sinks to slowly revert, in the absence of observational data, to "first guesses" of close to zero net annual mean for ecosystems. This procedure may have produced a bias. Also due to the sparseness of measurements, we had to assume coherence of ecosystem processes over large distances, giving existing observations perhaps an undue amount of weight. The process model for terrestrial photosynthesis and respiration was very "basic", and will likely be greatly improved in future releases of CarbonTracker. Easily the largest single annual mean source of CO2 is emissions from fossil fuel burning, which are currently not estimated by CarbonTracker. We use estimates from emissions inventories (economic accounting) and subtract those from the total sources derived by CarbonTracker. A small relative error in the inventories would thus translate into a larger relative error in the annual mean ecosystem sources/sinks that have smaller magnitudes. We expect to add a process model of fossil fuel combustion in future releases of CarbonTracker. Finally, additional measurement sites are expected to lead to the greatest improvements, especially to more credible and specific source/sink results at smaller spatial scales.

CarbonTracker is a NOAA contribution to the North American Carbon Program