Assessing lightweight satellite inversion methods for industrial NOx emissions in Spain

Accurate monitoring and estimation of pollutant emissions are essential for achieving global emission reduction commitments. High-point-source Nitrogen Oxides (NOx) primarily arise from combustion in power plants, cement facilities, petrochemical complexes, steel mills, and refineries. Traditional satellite-based top-down emission estimates rely on computationally intensive inversions using Chemical Transport Models (CTMs) that assimilate atmospheric composition data. However, recent lightweight inversion approaches provide an alternative that resolves emissions from individual sources with markedly reduced computational demand. In this study, we combine tropospheric NO₂ columns from the TROPOMI instrument on Sentinel-5 Precursor satellite with ERA5 wind fields over Spain with the open-source Python library ddeq v1.0 to estimate NOx emissions in 2021 from twenty large industrial sources at daily resolution. According to the Spain ministry inventory for 2021, recently developed from the HERMESΔ model, these twenty sites represent the strongest NOx industrial emitters in the Spanish peninsular and insular domain. We assess five lightweight point-source inversion techniques: Gaussian Plume (GP), Integrated Mass Enhancement (IME), Cross-Sectional Flux (CSF), Lightweight Cross-Sectional Flux (LCSF) and Flux Divergence (FD). Emission estimates are compared against the HERMESΔ emission model using metrics such as mean fractional bias and relative difference. Additionally, we have incorporated time-varying NOx lifetimes, derived from the CAMS EAC4 reanalysis to improve the accuracy of the emission estimates and, proposed and applied, explicit criteria for detecting scenes where the plumes can provide useful information to the inversions. For sources in the Canary Island, the lightweight inversions reproduce HERMESΔ emissions with smaller relative differences and tighter agreement than for mainland sources, which experience more complex flow and source interference. The results delineate the range of conditions where lightweight inversions deliver robust constraints on industrial NOx emissions in a low-to-moderate emission regime and they outline residual biases that motivate further development of lifetime parametrizations, plume detection criteria and inventory–satellite comparison strategies.