Deriving tropospheric NO2 layer heights from TROPOMI observations

Abstract:

Satellite observations of tropospheric NO₂ columns are being used in many applications from air quality studies to emission estimates. While the NO₂ absorption signal is relatively strong and easy to extract, quantification of the light path length and thus the sensitivity of the measurements is less accurate. As the vertical sensitivity of the measurements varies strongly in the lower atmosphere, a priori assumptions are needed on the NO₂ vertical distribution to convert the observed slant path absorptions to vertical column amounts. These a priori profiles are usually taken from atmospheric model runs.

The TROPOMI instrument, launched in October 2017 on the European Sentinel 5 precursor satellite, provides measurements at spatial resolution of up to 3.5 x 5.5 km². At these scales, it is challenging to provide appropriate a priori NO₂ profiles as regional models having comparable resolution are difficult to run on a global scale. As a result, often low resolution a rpiori data is used which can become a limitation for the accuracy of tropospheric NO₂ products from high-resolution satellite instruments.

In this study, we evaluate the potential to use the wavelength dependence of Rayleigh scattering for extracting information on the vertical distribution of the observed NO₂ directly from the satellite spectra. Sensitivity studies on synthetic spectra indicate that a simple semi-empirical approach can be used to determine an effective NO₂ layer height parameter, which can be used to select an appropriate a priori profile for the air mass factor calculation without the need for model data.

Application of the method to TROPOMI measurements demonstrates that qualitatively, situations with NO₂ in the boundary layer can be identified and distinguished from measurements where NO₂ is found in elevated layers. This proves that the method works in principle. A quantitative application is however limited to scenes with large NO₂ signals as uncertainties in the determination of the layer height propagate to the derived air mass factors and therby to teh tropospheric columns.