Overall, there has been very good progress in the first 15 months of the project.
WP1 has initiated the collection, processing and analysis of CO2, NOx, and CO emission factor (EF) data and their associated uncertainty at the national level for the road transport sector. WP1 also produced a collection of daily (day-of-the-year) emission temporal profiles and associated uncertainties for a selection of anthropogenic sectors. These profiles have been implemented in an Fossil Fuel Data Assimilation System (FFDAS) to produce a 4-dimensional CO2 emission ensemble, which will be used to describe uncertainty covariances for the global CO2MVS system. In addition, the global point source database constructed in the framework of the CoCO2 project was further developed resulting in a new version of the catalogue that updates the reference year (from 2018 to 2021), refines the emission ratios considered for power plant emissions (NOx:CO2, CO:CO2) and introduces cement and iron and steel plants. Finally, WP1 completed several tasks towards the preparation of the global CO2, NOx and CO emission uncertainty information needed for the modelling and testing of the IFS prior error covariance matrix performed in WP2. A first version of global gridded uncertainties for CO2, CO and NOx emissions and error correlations was produced to be tested in inversion experiments in WP2.
WP2 has implemented a peak-finding algorithm that was applied to TROPOMI NO2 observations on the Sentinel-5P satellite. The algorithm is now being applied to geostationary GEMS NO2 observations. Results were published in Deliverable 2.1. Data-driven methods for hot-spot emission quantification were developed for quantifying CO and NOx emissions from TROPOMI data and applied for quantifying emissions of hot spots in Africa and Europe. Simplified NOx chemistry schemes are also under development. For local inversions, a novel NO2-to-NOx conversion model was developed using high-resolution NOx simulations from the CoCO2 project. For the global system, full-chemistry simulations were conducted and used for training machine-learning models that predict NOx:NO2 ratios and NOx lifetimes. Finally, a data pipeline has been developed for ECMWF's IFS system that can be used to model the prior emission error covariance matrix (B) from the information collected in WP1.
The activities in WP3 were mainly focused on compiling databases of global background observations and flux maps for 14CO2 and APO. A database of the globally distributed background 14CO2 observations from four laboratories which have conducted long-term samplings worldwide was compiled and documented in deliverable D3.1. The report includes an extensive comparison of the datasets from the different laboratories to characterize the overall consistency of the database and the typical uncertainties in the observation time series. The production of the global database of APO observations was delayed due to the dependency on organisations outside the CORSO consortium. Mitigation actions have been activated to minimize the impact on other activities within the project. The flux maps database will be used as the prior estimates in the Bayesian framework of the atmospheric inverse models later in the project. WP3 also prepared the intensive observation sampling in Europe for the year 2024, which was started in January 2024. The different inverse modelling groups in WP3 also used the first year of the project to prepare their modelling configurations in preparation for further activities.
WP4 has made good progress with the development of neural network (NN) observation operators for ASCAT backscatter observations as well as satellite SIF observations. Results look promising but will need further fine-tuning. The impact of these additional observations to constrain the land surface model are being tested in the Météo-France data assimilation system, while at the same time they are being implemented in the Integrated Forecasting System (IFS) of ECMWF, which will be the core of the global system of the CAMS CO2MVS.
