Impacts of TROPOMI-Derived NOX Emissions on NO2 and O3 Simulations in the NCP during COVID-19

NO2 and O-3 simulations have great uncertainties during the COVID-19 epidemic, but their biases and spatial distributions can be improved with NO2 assimilations. This study adopted two top-down NOX inversions and estimated their impacts on NO2 and O-3 simulation for three periods: the normal operation period (P1), the epidemic lockdown period following the Spring Festival (P2), and back to work period (P3) in the North China Plain (NCP). Two TROPOspheric Monitoring Instrument (TROPOMI) NO2 retrievals came from the Royal Netherlands Meteorological Institute (KNMI) and the University of Science and Technology of China (USTC), respectively. Compared to the prior NOX emissions, the two TROPOMI posteriors greatly reduced the biases between simulations with in situ measurements (NO2 MREs: prior 85%, KNMI -27%, USTC -15%; O-3 MREs: Prior -39%, KNMI 18%, USTC 11%). The NOX budgets from the USTC posterior were 17-31% higher than those from the KNMI one. Consequently, surface NO2 levels constrained by USTCTROPOMI were 9-20% higher than those by the KNMI one, and O-3 is 6-12% lower. Moreover, USTC posterior simulations showed more significant changes in adjacent periods (surface NO2: P2 vs P1, -46%, P3 vs P2, +25%; surface O-3: P2 vs P1, +75%, P3 vs P2, +18%) than the KNMI one. For the transport flux in Beijing (BJ), the O-3 flux differed by 5-6% between the two posteriori simulations, but the difference of NO2 flux between P2 and P3 was significant, where the USTC posterior NO2 flux was 1.5-2 times higher than the KNMI one. Overall, our results highlight the discrepancies in NO2 and O-3 simulations constrained by two TROPOMI products and demonstrate that the USTC posterior has lower bias in the NCP during COVD-19.

Automated summary

This study aims to improve the simulations of NO2 and O-3 during the COVID-19 epidemic using NO2 assimilations. The results show discrepancies in the simulations constrained by two TROPOMI products (KNMI and USTC), with USTC posterior having lower bias.