Nitrogen dioxide decline and rebound observed by GOME-2 and TROPOMI during COVID-19 pandemic.

Since its first confirmed case in December 2019, coronavirus disease 2019 (COVID-19) has become a worldwide pandemic with more than 90 million confirmed cases by January 2021. Countries around the world have enforced lockdown measures to prevent the spread of the virus, introducing a temporal change of air pollutants such as nitrogen dioxide (NO2) that are strongly related to transportation, industry, and energy. In this study, NO2 variations over regions with strong responses to COVID-19 are analysed using datasets from the Global Ozone Monitoring Experiment-2 (GOME-2) sensor aboard the EUMETSAT Metop satellites and TROPOspheric Monitoring Instrument (TROPOMI) aboard the EU/ESA Sentinel-5 Precursor satellite. The global GOME-2 and TROPOMI NO2 datasets are generated at the German Aerospace Center (DLR) using harmonized retrieval algorithms; potential influences of the long-term trend and seasonal cycle, as well as the short-term meteorological variation, are taken into account statistically. We present the application of the GOME-2 data to analyze the lockdown-related NO2 variations for morning conditions. Consistent NO2 variations are observed for the GOME-2 measurements and the early afternoon TROPOMI data: regions with strong social responses to COVID-19 in Asia, Europe, North America, and South America show strong NO2 reductions of ∼ 30-50% on average due to restriction of social and economic activities, followed by a gradual rebound with lifted restriction measures. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s11869-021-01046-2.

Applications of neural network methods to the processing of earth observation satellite data.

The new generation of earth observation satellites carries advanced sensors that will gather very precise data for studying the Earth system and global climate. This paper shows that neural network methods can be successfully used for solving forward and inverse remote sensing problems, providing both accurate and fast solutions. Two examples of multi-neural network systems for the determination of cloud properties and for the retrieval of total columns of ozone using satellite data are presented. The developed algorithms based on multi-neural network are currently being used for the operational processing of European atmospheric satellite sensors and will play a key role in related satellite missions planed for the near future.