Long-term variation in the tropospheric nitrogen dioxide vertical column density over Korea and Japan from the MAX-DOAS network from 2007 to 2017

Abstract:

As a part of the Multi-Axis Differential optical absorption spectroscopy (MAX-DOAS) network in Russia and ASia (MADRAS), instruments have been deployed at key locations, including polluted urban and clean remote sites, since 2007. We investigated the long-term observations of the tropospheric nitrogen dioxide vertical column density (NO₂ TropVCD) from MADRAS till 2017 at urban (Yokosuka and Gwangju) and remote (Fukue and Cape Hedo) sites in East Asia. The retrieval products, aerosol optical depth (AOD) at 476 nm and NO₂ TropVCD, using the DOAS technique and optimal estimation method were compared with co-located sun/sky radiometers and the Ozone Monitoring Instrument (OMI; OMNO2d), respectively. The overall correlation coefficients of monthly mean AOD (except at Yokosuka) and the NO₂ TropVCD between MAX-DOAS and other instruments were sufficiently high, at 0.79 and 0.95, but the monthly mean AOD from sun/sky radiometers and NO₂ TropVCD from OMI were lower than those from MAX-DOAS by 8% and 27%, respectively, due to the different measurement methods, uncertainty in a scale factor of the O₄ cross-section (for AOD), and uncertainty in the air mass factor (for NO₂ TropVCD). The level of monthly variation in the NO₂ TropVCD from MAX-DOAS matched to the OMI overpass time (~13:00 local time) showed good agreement with the OMI data during the summer but the difference between the two dataset increased in winter because of the homogenous photochemical dissociation of NO₂ in summer and the inhomogeneous spatial distribution of NO₂ in winter, respectively. The Theil-Sen slope of the long-term trend indicated a gradual reduction in NO₂ at all the Japanese sites, regardless of the season but no significant change in NO₂ at Gwangju, South Korea. To the contrary, the OMI satellite data revealed an increase in NO₂ TropVCD at all sites except for Yokosuka, where its similar decrease trend with MAX-DOAS, suggesting that results from satellite could be inappropriate to investigate trend analysis in less polluted or remote areas. A typical double peak in the diurnal cycle of the NO₂ VCD at less than 1 km was observed in the urban area during the summer, but the peak in winter occurred in the early evening instead of the early morning due to the longer lifetime of NO₂ and the meteorological conditions in winter. By using the backward trajectory, the identified potential source regions at the urban sites pointed out the emission hotspots in the Regional Emission inventory in ASia (REAS), but the spatial distribution from the OMI data showed good agreement with the potential source regions at only Yokosuka because there was a continuous influence from the Tokyo metropolitan area, where the NO₂ TropVCD measured by simultaneous satellite observations was also high. Similarly, the disagreement of spatial distribution between the OMI and potential source region in Gwangju was attributable to the lower influence from the nearby hotspot area, Seoul.