Applying a high-resolution atmospheric inversion framework to CO2 observations using GRAMM/GRAL

As urban areas encompass more than 70% of anthropogenic CO₂ emissions, they are a crucial target for effective climate change mitigation. To monitor and verify CO₂ mitigation strategies, atmospheric CO₂ measurements can be assimilated into a Bayesian inversion framework to infer CO₂ fluxes. Bayesian inversions combine knowledge of prior fluxes with atmospheric measurements under consideration of the atmospheric transport and its associated uncertainties to obtain CO₂ fluxes. It is particularly interesting to constrain sub-urban and sector-specific CO₂ fluxes for urban areas to pinpoint relevant emissions and to frame specific mitigation strategies. 

However, this approach requires the simulation of atmospheric transport processes at high-resolution to account for urban geometries and heterogeneities in emission patterns. We, therefore, use the GRAMM/GRAL model to simulate the atmospheric transport. GRAMM/GRAL computes concentration fields at high-resolution (e.g. 10x10m) using the Reynolds-Averaged Navier-Stokes equations together with a catalogue approach to pre-compute a set of meteorological situations (May et al., 2023). The catalogue approach allows the simulation of high-resolution concentration fields over long time periods because the computation time is decoupled from the simulation time. Therefore, the model is well suited for synthetic inversion studies conducted within an observing system simulation experiment (OSSE).

We developed a framework for high-resolution observing system simulation experiments (OSSE's) to analyse different urban network configurations with varying number, precision and location of sensors. We tested this framework over the city of Heidelberg, Germany, and evaluated different potential measurement network configurations to constrain the fossil fuel CO₂ emissions (Vardag and Maiwald, 2023). Building on this development, we now seek to apply this framework to the Paris metropolitan area, where an actual CO₂ measurement network has already been deployed (Horizon Europe, PAUL project). The network consists of multiple sensors of different types, which allows us to analyse different subsets of the sensors and compare their performances. We will present an outlook on the capabilities and shortcomings of our high-resolution inversion framework using the actual measurement network to estimate CO₂ emissions. The results will provide insight on possible measurement network improvements, as well as on technical improvements of the framework.

May, Maximilian, Simone Wald, Ivo Suter, Dominik Brunner, and Sanam N. Vardag. 2024. “Evaluation of the GRAMM/GRAL Model for High-Resolution Wind Fields in Heidelberg, Germany.” Atmospheric Research 300 (April): 107207. https://doi.org/10.1016/j.atmosres.2023.107207.

Vardag, Sanam N., and Robert Maiwald. 2023. “Optimising Urban Measurement Networks for CO2 Flux Estimation: A High-Resolution Observing System Simulation Experiment Using GRAMM/GRAL.” Geoscientific Model Development Discussions, October, 1–28. https://doi.org/10.5194/gmd-2023-192.