The COVID-19 pandemic led to widespread lockdowns, causing a global reduction in air pollutant emissions. Previous studies using high-resolution instruments like TROPOMI and OMI have investigated these changes, but data from Chinese satellite instruments have been lacking due to limitations in spectral quality. The EMI instrument onboard the GaoFen-5 satellite, the first Chinese satellite-based ultraviolet-visible hyperspectral spectrometer, presents a unique opportunity to address this gap. This study aimed to overcome the challenges posed by the lower spectral quality of EMI and retrieve global gaseous pollutants to analyze air quality variations during the pandemic. The ability to compare air quality across different countries and regions using satellite data provides a significant advantage over ground-based observations due to its improved spatial coverage and data consistency. The study focuses on analyzing the variations in NO2, SO2, and HCHO to understand the impact of lockdown measures on air pollution and its connection to economic factors.

Numerous studies have used space-borne observations from high-resolution instruments like TROPOMI and OMI to analyze the reduction in atmospheric nitrogen dioxide (NO2) and other pollutants during the COVID-19 pandemic. These studies have leveraged in-situ monitoring and atmospheric chemical transport modeling to complement satellite data. Bauwens et al. and Sun et al. are examples of studies that explored global changes in NO2 and formaldehyde (HCHO) levels using satellite observations. However, prior to this study, no research had utilized data from Chinese satellite instruments, primarily due to the limitations in the spectral quality of the available instruments. This study aimed to fill this knowledge gap by utilizing data from EMI, after significant improvements in data retrieval techniques.

The study used data from the EMI instrument onboard the GaoFen-5 satellite, a push-broom spectrometer covering ultraviolet and visible spectral bands (240–710 nm) with a nadir spatial resolution of 12 × 13 km². The retrieval of NO2, SO2, and HCHO TVCDs involved several steps and algorithm optimizations to address the challenges posed by the inferior spectral quality of EMI compared to TROPOMI. For NO2, the retrieval involved three steps: (1) retrieval of total NO2 slant column density (SCD) using the differential optical absorption spectroscopy (DOAS) technique, considering absorption cross-sections of various gases; (2) obtaining tropospheric NO2 SCD after stratospheric NO2 separation using the STRatospheric Estimation Algorithm from Mainz; and (3) converting tropospheric SCDs to tropospheric VCDs (TVCDs) using air mass factors (AMFs). HCHO retrieval involved similar steps using the basic optical differential spectroscopy (BOAS) method and reference sector correction. SO2 TVCDs were retrieved using the optimal estimation (OE) algorithm, employing the vector linearized discrete ordinate radiative transfer model (VLIDORT). To improve the quality of EMI retrievals, several novel methods were developed. These included on-orbit wavelength calibration to account for changes in the full width at half maximum (FWHM) of the instrumental spectral response functions (ISRFs) and wavelength shifts. An adaptive iterative retrieval algorithm was implemented to minimize interference from other trace gases and improve the signal-to-noise ratio (SNR). Finally, simulated irradiance was used in place of measured irradiance due to limitations in the EMI instrument's solar spectrum measurements. The retrieved data were validated using surface observations and comparisons with TROPOMI data.

The study's key findings include: (1) Successful retrieval of NO2, SO2, and HCHO TVCDs from EMI data despite its initially poor spectral quality, showcasing the effectiveness of the developed algorithm optimizations. (2) A significant (20%) decrease in global average NO2 TVCD in March 2020 compared to March 2019, with more pronounced reductions in regions with high concentrations (eastern China, western Europe, and eastern North America). The timing of NO2 decreases strongly correlated with the implementation of lockdown measures in various cities. (3) A slight global increase in SO2 TVCD in March 2020 relative to March 2019, with no significant reduction in regions with substantial SO2 emissions like India, suggesting that these emissions were less affected by the pandemic at least initially. (4) A 21% decrease in global average HCHO TVCD in March 2020 compared to March 2019, showing regional variations that indicated different dominance between anthropogenic and natural sources of VOCs. (5) Good consistency was observed between monthly NO2 and HCHO TVCDs from EMI and TROPOMI, validating the EMI retrieval results. (6) Analysis comparing NO2 variations with GDP growth revealed differential effects of the pandemic on the primary, secondary, and tertiary industrial sectors in different countries.

The study successfully demonstrated the utility of EMI data for studying global air quality changes during the COVID-19 pandemic, despite its initial limitations. The observed decrease in NO2 levels directly correlated with lockdown measures, providing strong evidence for the significant contribution of anthropogenic emissions to NO2 concentrations. The lack of a substantial decrease in SO2, particularly in India, highlighted the relative resilience of some emission sources (e.g., power generation) to pandemic-related restrictions. The variability in HCHO reductions provided insights into the diverse sources of VOCs in different regions. The comparison of NO2 changes with GDP trends underscored the heterogenous economic consequences of the pandemic across countries and economic sectors. The results confirm the potential of advanced data processing techniques to extract meaningful information from even lower-quality satellite data and enhance our understanding of the complex interplay between human activity, environmental regulations, and air quality.

This study successfully retrieved global gaseous pollutants from the EMI instrument, overcoming initial challenges related to its spectral quality. The analysis provided valuable insights into the impact of the COVID-19 pandemic on air quality across different regions and economic sectors. Future research could focus on improving the retrieval algorithms for EMI, extending the analysis to other pollutants, and exploring the long-term implications of pandemic-related emission reductions on air quality and climate change. The success of this study highlights the potential of combining advanced data processing techniques with existing datasets to extract valuable information and enhance our understanding of complex environmental phenomena.

The study acknowledged limitations related to the initial spectral quality of the EMI instrument, requiring significant algorithm optimizations. The retrieval of SO2 presented greater challenges due to noise and the proximity of the strongest absorption band to the spectral edge, potentially affecting the accuracy of the SO2 data. The reliance on TROPOMI data for certain parameters in AMF calculations might have introduced some systematic errors. Furthermore, the study's focus was on March 2020, and further investigation is needed to fully assess the long-term impacts of the pandemic on air quality.