Ground-level NO2 concentration estimation based on OMI tropospheric NO2 and its spatiotemporal characteristics in typical regions of China

Although the ground-level NO2 measurement from air quality monitoring sites is relatively accurate, it a challenge to obtain continuous spatial coverage due discrete distribution of sites. Thus, tropospheric column amount satellites with wide and temporal higher resolution has been increasingly used estimate NO2. However, most estimation methods were performed spontaneously using simple linear model throughout study period. These simplified models improve efficiency development enhance generality application, but they ignore fact that contributors changes are not always consistent time. This considered fixed random effects influencing factors developed mixed effect (MEM) By data in China January 1, 2014 June 30, 2020 other multivariate auxiliary such as meteorological elements terrain elevation, reliability daily typical populated areas estimated by MEM was evaluated. The average monthly R2 10-fold CV each area during 2014–2020 greater than 0.60 proportion 0.7 about 71%, suggesting MEM. It found characteristics more distinct, influential also different. In addition, associated control policies emission reduction measures various regions, shown an overall downward trend 2014–2019. uncertainty daily-scale boundary layer conditions can lead varying degrees deviations predictions Validation station observations demonstrates prediction at seasonal time scale (R2 = 0.81, RMSE 3.86 μg/m3) performs better those scales 0.65, 0.75, 7.92, 6.24 μg/m3). Therefore, method averaging be accuracy on individual dates. summary, this shows promising modeling method, effective for pollution mapping large geographic region.