ML-based multi-regional sensitivity assessment of ship NO₂ plumes detection using TROPOMI data

This study applies machine learning-based methods to investigate the sensitivity limits of a detection system for NO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> plumes from seagoing ships using TROPOMI data. By combining datasets from four regions located on the Europe – Asia trade route, we study the consistency of the results reported in a previous study. Findings reveal that sensitivity limits established for the Arabian and Mediterranean Seas remain relevant when merging data from all four regions. Moreover, the combined dataset enhances the quality of plume recognition from the strongest emitters and amplifies the impact of crucial features on model outcomes.