Improved Anthropogenic SO2 Retrieval from High-Spatial-Resolution Satellite and its Application during the COVID-19 Pandemic

Sulfur dioxide (SO2) measured by satellites is widely used to estimate anthropogenic emissions. The Sentinel-5 Precursor (S-5P) operational SO2 product is overestimated compared to the ground-based multiaxis differential optical absorption spectroscopy (MAX-DOAS) measurements in China and shows an opposite variation to the surface measurements, which limits the application of TROPOspheric monitoring instrument (TROPOMI) products in emissions research. Radiometric calibration, a priori profiles, and fitting windows might cause the overestimation of S-5P operational SO2 product. Here, we improve the optimal-estimation-based algorithm through several calibration methods. The improved retrieval agrees reasonably well with the ground-based measurements (R > 0.70, bias <13.7%) and has smaller biases (-28.9%) with surface measurements over China and India. It revealed that the SO2 column in March 2020 decreased by 51.6% compared to March 2019 due to the lockdown for curbing the spread of the COVID-19 pandemic, and there was a decrease of 50% during the lockdown than those after the lockdown, similar to the surface measurement trend, while S-5P operational SO2 product showed an unrealistic increase of 19%. In India, the improved retrieval identified obvious hot spots and observed a 30% decrease of SO2 columns during the lockdown.

Automated summary

Satellite measurements of sulfur dioxide (SO2) are widely used to estimate anthropogenic emissions, but the operational SO2 product of Sentinel-5 Precursor (S-5P) has been found to be overestimated compared to ground-based measurements in China and India. By improving the algorithm, the updated retrieval shows better agreement with ground-based measurements and reveals a significant decrease in SO2 emissions during the COVID-19 lockdown period.