This document reports on the work and the results obtained in task 3.5 ‘Perform QND experiments with an advanced data assimilation system (CCFFDAS) to establish inversion strategy’. It provides a description of the CCFFDAS modelling framework, the applied QND methodology as well as the results obtained from the work performed in task 3.5 ‘Perform QND experiments with an advanced data assimilation system (CCFFDAS) to establish inversion strategy’. We have developed a prototype of a global Carbon Cycle Fossil Fuel Data Assimilation System (CCFFDAS) that combines the functionalities of the Fossil Fuel Data Assimilation System by Asefi-Najafabady et al. (2014) and of the Carbon Cycle Data Assimilation System by Kaminski et al. (2017a). The CCFFDAS was applied to assess a number of design aspects of the upcoming Monitoring and Verification Support capacity as part of the Copernicus CO2 Monitoring mission (CO2M). The assessment was based on the Quantitative Network Design technique and quantified the mission’s performance in terms of the posterior uncertainty in the totals of the sectorial fossil CO2 emission rates for selected countries and the first week in June 2008. The emissions were classified into two sectors, one for electricity generation and the other for all other emissions denoted as the “other” sector. We analysed two different observing networks, ground based in situ observations and satellite based total column observations, in a range of configurations.
We find that
- At country scale, a single CO2M satellite achieves posterior uncertainties in the weekly sectorial emission rates well below the weekly emission rate for the national total for the sectors, assuming that annual emissions are evenly divided between all weeks.
- For each sector, a constellation of four satellites achieves a larger reduction in posterior uncertainty, i.e. the number of satellites matters. This added value of the extra satellites is more pronounced for the other sector than the electricity generation sector, where the uncertainty over China is reduced by about 30% for this particular case (first week in June 2008) when going from a single satellite to the four-satellite constellation. This is below the theoretical possible 50% reduction due to the different satellite tracks and with it differences in cloud cover and systematic errors for this particular week.
- Correlated systematic errors in the satellite XCO2 can seriously degrade the performance, resulting in increased posterior uncertainties from 15% to almost 130% relative to the case with no systematic errors.
- The representation error due to the mismatch in resolution between the satellite images and the atmospheric transport model increases posterior uncertainties, however, these uncertainties will be decreased when the resolution of the transport model is increased.
- If national inventories at weekly time scale were available, their inclusion would result in a strong reduction in posterior uncertainty (the largest reduction we obtained from all our experiments) compared with using atmospheric observations only, which quantifies the synergy / complementarity between the atmospheric (and other) observations and the inventories. The synergistic exploitation of diverse data steams is a particular strength of the CCFFDAS approach. However, by including national inventory data in the assimilation the obtained results are not an independent estimation of the emissions from atmospheric data anymore.
- A hypothetical projection of the posterior uncertainties in the weekly emission rate to the annual scale shows the potential for a CCFFDAS / CO2M verification mode (operating independently from the inventories), which may provide useful information that is complementary to the inventories with uncertainty ranges in the same order of magnitude.
- The inclusion of surface observations does not result in substantial uncertainty reductions for sectorial emissions compared with the default experiment with one satellite. However, we note that the current in situ network was not designed for observing anthropogenic emissions, nevertheless it is important to observe background conditions and validate satellite measurements.
- Adding radiocarbon observations helps to further constrain emission rates from the other sector especially for countries with a larger proportion of biogenic fluxes such as e.g. Brazil and Poland.