French NOx emissions as estimated from TROPOMI-PAL NO2 observations

Since 2018, TROPOMI onboard Sentinel-5P has brought images of NO2 tropospheric columns at an unprecedented high spatial resolution. We attempt to exploit the potential of this high-resolution information to estimate NOx emissions at the national scale based on atmospheric inversion approaches, with a 10-km spatial resolution. This study focuses on France and assesses the Covid-19 pandemic effects.Our analysis is based on the variational mode of the recently developed Community Inversion Framework (CIF), coupled with France's configuration of the CHIMERE regional chemistry transport model and its adjoint code. Both CHIMERE and its adjoint code include the MELCHIOR-2 chemical scheme with more than 100 reactions (24 for inorganic chemistry). This variational framework allows to solve high-dimensional inversion problems and thus investigate fine-scale patterns in the concentration and emission fields. It also allows to properly account for the non-linearities associated with the chemistry. CHIMERE is driven by the European Centre for Medium-Range Weather Forecasts (ECMWF) operational meteorological forecast. The inversion corrects anthropogenic emissions from the French National Spatialized Inventory (INS, Inventaire National Spatialisé), mapping NOx emissions at a 1 x 1 km2 horizontal resolution. It also corrects biogenic emissions from the MEGAN model. We use the last TROPOMI reprocessed data: the PAL product. We present different tests that have been conducted to improve the implementation of the observation and control vector in order to strengthen the robustness of the inversion. Then, we evaluate the potential of the TROPOMI-PAL observations to quantify emissions at the national to local scale and at the annual to monthly scale based on the analysis of the reference inversions for the years 2019 and 2020. Finally, we focus on the differences found between the emission estimates for spring 2020 and 2019 to assess the impact of the Covid-19 pandemic.