Comparisons of Changes and Long-term Trends of Tropospheric NO2 Vertical Column Densities from Different Products over China from 2005 to 2019

Abstract:

Nitrogen oxides are one of the most important air pollutants in the atmosphere. Although satellite-based tropospheric NO₂ VCDs retrieval provides abundant information for research fields including NO_x emission retrieval, atmospheric chemistry, global/region air pollution and long-term trend analysis, errors in the retrieval algorithms from different products can result in significantly different conclusions, especially from the treatments of aerosol optical effects, surface reflectance and priori NO₂ profiles. Here, we present the comparisons of changes and long-term trends of tropospheric NO₂ VCDs over China from 2005 to 2019, retrieved from POMINO v1.1 and QA4ECV. POMINO v1.1 retrieval accounts for the anisotropy of surface reflectance using BRDF with explicit aerosol corrections and higher-resolution profiles, while QA4ECV uses LER approximation with implicit aerosol corrections and relatively coarse-resolution profiles. Compared with explicit corrections, implicit aerosol corrections that simply treat aerosols as “effective clouds” tend to increase cloud fraction and decrease tropospheric NO₂ AMF simultaneously, leading to 10% to 30% underestimation of tropospheric NO₂ VCDs. Furthermore, higher cloud fraction caused by implicit corrections can exclude pixels under heavy pollution during sampling, which contributes to the differences between products combined with different sampling principles. Besides, Retrieved VCDs can be 5% to 15% lower due to LER’s isotropic approximations, and higher-resolution priori NO₂ profiles can reveal details of small NO₂ hotspots and spatial gradient around them, which are better for regional trend analysis than coarse-resolution profiles.