Validation of OMI-DOAS total ozone column amounts against ground-based measurements at an African equatorial belt site

Appl Opt. 2020 Nov 1;59(31):9896-9904. doi: 10.1364/AO.396764.

Abstract

The total ozone column amount (TOCA) values from the Ozone Monitoring Instrument (OMI) derived from OMI/Aura ozone (O3) differential optical absorption spectroscopy (DOAS) V003 (OMDOAO3) have been validated against the ground-based TOCA values derived from Dobson and the Norwegian Institute for Air Research UV measurements in Kampala (0.31º N, 32.58º E, 1200 m), Uganda, for the period between 2005 and 2018. Under all-sky conditions, the OMI retrieval algorithm was found to underestimate the TOCA values with mean bias (MnB), root mean square error (RMSE), and correlation coefficient (r) values ranging from about -3.4% to -1.7%, 2.4% to 4.9%, and 0.73 to 0.90, respectively. When only days with a radiation modification factor greater than or equal to 65% were considered, the MnB, RMSE, and r values between TOCA values derived from ground-based and OMI measurements improved, and they ranged from -2.5% to -1.3%, 1.4% to 3.8%, and 0.8 to 0.91, respectively. A good agreement was found between TOCA values derived from Dobson measurements and those derived from OMI satellite measurements with MnB, RMSE, and r values of about -1.8%, 1.4%, and 0.91, respectively. This was due to the fact that Dobson measurements were taken only when the sky was perceived clear. The underestimation of TOCA values by the OMI retrieval algorithm was found to be due mainly to clouds and aerosols.