Evaluating the impact of using Δ14CO2 observations in quantifying fossil CO2 emissions over Europe

CO₂ is one of the most important greenhouse gases responsible for global climate change. Today’s atmospheric CO₂ concentration has risen by nearly 49% compared to pre-industrial times, mainly caused by fossil fuel combustion and cement production. CO₂ emission reports under the Paris Agreement are crucial in understanding the main sources responsible for these emissions, and spatial and temporal distributed emission inventories such as EDGAR and ODIAC have been important tools to gain additional insights about when and where these emissions happen. However, increasing the resolution comes with an increment in the uncertainty at sub-annual and sub-national scales. One way to improve the knowledge on these emission inventories is by inverting in situ observations of CO₂ and Δ¹⁴CO₂ at regional scales. Radiocarbon (¹⁴C) is the radioactive isotope of carbon, and due to its half-life time of ~5730 years, it is not present in fossil fuels, making it a good tracer for the natural carbon cycle.

In this study, we evaluate the impact of incorporating Δ¹⁴CO₂ observations from the Integrated Carbon Observation System (ICOS) network and its sampling strategy over Europe into the Lund University Modular Inverse Algorithm (LUMIA) for optimizing fossil CO₂ and the biosphere fluxes in a horizontal grid of 0.5° x 0.5° and on a weekly temporal resolution. Δ¹⁴CO₂ is currently mostly measured in 2-weekly integrated samples. As part of the EU-funded CORSO (CO2MVS Research on Supplementary Observations) Project, an intensive campaign collecting 1-hour flask samples every third day at 10 Western/Central European sampling stations will be performed during 2024. W perform a series of Observing System Simulation Experiments (OSSEs) using various model products (EDGAR/ODIAC, LPJ-GUESS/VPRM) as prior fluxes and as assumed true fluxes for fossil and biosphere fluxes to generate a time series of synthetic observations. We first demonstrate the impact of adding Δ¹⁴CO₂ observations in addition to the CO₂ observations to recover the true fossil and biosphere flux time series and the assumed true total CO₂ annual budget over Europe. Further, we assess the impact of the sampling strategy by comparing a simulation using only Δ¹⁴CO₂ integrated samples against a simulation including the CORSO sampling strategy. In the latter case, we find a notable improvement in recovering the fossil CO₂ emissions in Western/Central Europe and countries such as Germany and France. We also evaluate other sampling strategies, such as selecting observations with the largest apportionment of fossil CO₂ and the lowest impact of nuclear emissions. Such approaches seem to be a promising way to improve the quantification of fossil CO₂ emissions in regions with a dense sampling network and neighbouring regions such as Eastern Europe.