Marina Zara

Satellite observations of NO2 provide unique insight into air quality with daily global coverage that ground-based networks cannot deliver. Instruments like the Ozone Monitoring Instrument (OMI) and the Global Ozone Monitoring Experiment-2 (GOME-2) have collected multi-decadal measurements essential for detecting trends in atmospheric composition. However, uncertainties in satellite retrievals - from spectral fitting errors, stratosphere-troposphere separation, air mass factor computation - can obscure emission changes and limit environmental policy effectiveness assessments. This thesis addresses such limitations and demonstrates the capability of high-quality satellite data to characterise emission patterns and chemical regime shifts. Reliable trend detection lies in minimizing retrieval uncertainties, which this thesis addresses by evaluating spectral fitting algorithms from the EU’s QA4ECV project against NASA’s operational products. Comprehensive intercomparison of OMI and GOME-2A NO2 and formaldehyde retrievals characterised uncertainty budgets and demonstrated improvements in random and systematic errors. Advanced corrections for stratospheric contributions, air mass factors, and temperature dependencies further reduce biases, establishing a quality-assured retrieval framework that forms the basis for investigating atmospheric composition changes with confidence. Applying these advanced retrievals to pollution hotspots like the Netherlands reveals striking NO2 reductions over the years, closely tracking ground-based measurements and emission inventories. Yet concurrent rising ozone concentrations signal shifts in atmospheric chemistry. Detailed spatial and seasonal analysis demonstrates satellites’ capability - under careful interpretation - to serve as robust proxies for underlying emission changes and their impact on atmospheric chemistry. Multi-species satellite observations extend beyond trend quantification to source characterisation. In rapidly-developing China, combining OMI NO2 and SO2 observations enables qualitative classification of emission source types - distinguishing industry, power generation, transportation, nature - based on their characteristic SO2:NOx signatures. Spatiotemporal emission source redistribution highlights the evolution of regional patterns in response to policy interventions, as point sources shift geographically. Advanced satellite retrievals reveal not only whether emissions change, but how and where they respond to environmental measures and socioeconomic activities. This thesis establishes satellite NO2 observations as essential tools for trend investigation, emission monitoring, air-quality policy assessment, and the understanding of atmospheric chemical evolution. The long-term, quality-assured data records from OMI and GOME-2A - now extended by TROPOMI with unprecedented spatial resolution - enable high-quality trend analyses critical for pursuing air quality and climate goals, strengthening our capacity to monitor Earth’s changing atmosphere from space.