Downward trend of NO2 in the urban areas of Beijing-Tianjin-Hebei region from 2014 to 2020: Comparison of satellite retrievals, ground observations, and emission inventories

Abstract

Significant downward trend of NO2 over BTH region has been detected by both satellite and ground surface observations in recent years due to the continuous regulatory policies in China. Previous studies mostly focused on the scientific issue of analyzing the distribution characteristics of NO2 and the drivers of interannual NO2 variations. Whether the declines derived from independent data sets with different spatiotemporal representations are comparable? If they are, what is the significance of these observed quantitative interannual trends for evaluating the estimated NOx emission changes is poorly characterized. This issue is still need to be brought to more attention.This study first investigates the variations of the long-term trends derived from satellite and ground observations, and then uses these data to evaluate the updates of NOx emissions from MEIC (Multi-resolution Emission Inventory for China) inventory during 2014–2020 over the BTH (Beijing-Tianjin-Hebei) region. In the 12 cities examined here, both the OMI (Ozone Monitoring Instrument) and the AQS (Air Quality System) detected comparable and substantial downward trends, with the overall reduction ranging from 28.2% to 49.5% and from 27.4% to 54.8% in urban areas according to OMI and AQS, respectively. The NOx emission adopted from MEIC showed a similar downward trend, but with a much lower reduction ranging from 4.3% to 32.5% (from 2014 to 2019). Large discrepancies were also found in the annual trends between NOx emissions (MEIC) and NO2 observations (OMI and AQS) especially after 2016, suggesting that the magnitude of recent emission reductions over the BTH region doesn't seem to be well captured by the MEIC emission projection, and in urban areas, in particular, there are significant overestimations. Assuming a linear relationship between NOx emissions and NO2 columns, 45.2% overestimation errors of simulations using priori NOx emissions can be explained by using posteriori inventory optimized by OMI observation. Our findings demonstrate that the combination of satellite and ground observations is likely to provide more reliable estimates of NOx emission and its trend over the BTH region under continuous emission reductions.