Comparing space-based to reported carbon monoxide emission estimates for Europe’s iron and steel plants.

Carbon monoxide (CO) is an air pollutant that plays an important role in atmospheric chemistry and its emissions can serve as a proxy for CO₂emissions. Within the European Union, large industrial emitters are required to report their annual facility-level emissions to the European Pollutant Release and Transfer Register (E-PRTR) based on stack measurements or calculations. We use TROPOMI satellite observations of carbon monoxide concentrations to estimate the emissions and compare these to the E-PRTR reports. Since 2018, the TROPOMI satellite instrument observes carbon monoxide concentrations at a resolution down to 5.5x7 km²with daily global coverage. CO plumes from large iron and steel plants are clearly visible in the TROPOMI data, enabling us to estimate the plants’ annual emissions. To achieve this, we perform high-resolution atmospheric transport simulations with the WRF model for 2019 over the 21 highest emitting iron and steel plants in Europe. We combine the simulations with TROPOMI observed concentrations in an analytical inversion to estimate the annual emission rates of the individual plants using the E-PRTR emissions as prior estimates. The TROPOMI-based emission estimates generally agree well with the reported emission rates (R² = 0.87) while showing limited sensitivity to the prior emission estimate. For 10 out of 21 plants the reported and TROPOMI-estimated emission rates agree within 20% whereas 6 plants show discrepancies over 40%. For the plants with the largest emissions, we perform an additional emission quantification using the Cross-Sectional Flux (CSF) method, which does not use any prior knowledge. These CSF estimates are consistent with the inversions, providing additional confidence in the space-based emission estimates. For two plants for which the 2019 inversion-based emission estimates are significantly different from the reported emission rates, we extend our analysis to 2020. The inversion estimates for 2020 agree with those from 2019 and match with reported emissions for 2020, raising questions on reported emissions for 2019. Our work shows that we can use the TROPOMI observations to reliably estimate CO emissions from large iron and steel plants and how these analyses can be used to identify uncertainties in reported emissions.