Underestimation of column NO2 amounts from the OMI satellite compared to diurnally varying ground-based retrievals from multiple PANDORA spectrometer instruments

NADER ABUHASSAN; MANVENDRA DUBEY; JHOON KIM; MARCELO RAPONI; JAY HERMAN; JAE KIM; MARIA TZORTZIOU

Atmospheric Measurement Techniques

Copernicus Publications

Año: 2019 vol. 12 p. 5593 - 5612

1867-1381

Retrievals of total column NO2 (TCNO2) are compared for 14 sitesfrom the Ozone Measuring Instrument (OMI using OMNO2-NASA v3.1) on the AURAsatellite and from multiple ground-based PANDORA spectrometer instrumentsmaking direct-sun measurements. While OMI accurately provides the dailyglobal distribution of retrieved TCNO2, OMI almost alwaysunderestimates the local amount of TCNO2 by 50 % to 100 % in pollutedareas, while occasionally the daily OMI value exceeds that measured byPANDORA at very clean sites. Compared to local ground-based or aircraftmeasurements, OMI cannot resolve spatially variable TCNO2 pollutionwithin a city or urban areas, which makes it less suitable for air qualityassessments related to human health. In addition to systematicunderestimates in polluted areas, OMI´s selected 13:30 Equator crossing timepolar orbit causes it to miss the frequently much higher values ofTCNO2 that occur before or after the OMI overpass time. Six discussedNorthern Hemisphere PANDORA sites have multi-year data records (Busan,Seoul, Washington DC, Waterflow, New Mexico, Boulder, Colorado, and Mauna Loa),and one site in the Southern Hemisphere (Buenos Aires, Argentina). The firstfour of these sites and Buenos Aires frequently have high TCNO2(TCNO2 > 0.5 DU). Eight additional sites have shorter-termdata records in the US and South Korea. One of these is a 1-year datarecord from a highly polluted site at City College in New York City withpollution levels comparable to Seoul, South Korea. OMI-estimated air massfactor, surface reflectivity, and the OMI 24 km × 13 km FOV (field of view)are three factors that can cause OMI to underestimate TCNO2. Because ofthe local inhomogeneity of NOx emissions, the large OMI FOV is the mostlikely factor for consistent underestimates when comparing OMI TCNO2 toretrievals from the small PANDORA effective FOV (measured in m2)calculated from the solar diameter of 0.5∘.