In order to make sure that the radiative transfer models MYSTIC (to be used to generate synthetic data) and LIDORT (used in BIRA's NO2 retrieval algorithm) are consistent, we first performed a clear sky model intercomparison.

Surface albedos: 0, 0.1, 0.3, 0.8

solar zenith angles [degrees]: 30,60

viewing zenith angles [degrees]: 0,30,60

viewing azimth angles [degrees]: 0,90,180

Rayleigh scattering parameterization:

MYSTIC uses Bodhaine 1999: Bodhaine, B. A., Wood, N. B., Dutton, E. G., and Slusser, J. R.: On Rayleigh optical depth calculations, J. Atmos. Ocean Tech., 16, 1854–1861, 1999.

Spectral range: 400-500nm (0.2 nm resolution)

Molecular atmosphere: US-standard with modified NO2 profiles

Solar irradiance: Solar irradiance data.

NO2 profiles: Ocean, low pollution, Pacific, medium pollution, European, high pollution

Absorption cross sections: NO2, O3, O4, H2O

Overall, we have compared spectra for 216 cases, results at 400 nm are shown in the following figure. MYSTIC results are shown as blue circles, LIDORT as red circles. Relative differences are below 1.5%.

The following figures show various spectra. In each column one parameter is varied to see the sensitivity. All other parameters are fixed and correspond to the values in the title of the figure.

MYSTIC Results:

Results for 400-500 nm

We find a good agreement between MYSTIC and LIDORT, remaining difference can be attributed to statistical error of Monte Carlo method.

We find mostly a good agreement for box AMFs.

Spectral range: 755-775nm (0.005 nm spectral resolution)

Molecular atmosphere: midlatitude-summer Midlatitude summer atmosphere

Absorption optical thickness Line-by-line absorption optical thickness

Solar irradiance Solar irradiance

The following figure shows the results and relative differences for 755nm, outside the O2A band. Good agreement (<1.5%).

The following figures show example spectra for the full range and for a zoom showing the ange between 767 and 768 nm.

Agreement between models quite good. Relative differences mostly below 1%. As expected, there are larger relative differences in line centers, where radiance becomes 0 or close to 0.

#Huan, should we compare the spectra convolved with instrument filter function?

MYSTIC Results: Results for 755-775 nm

Averaged values for 3 wavelengths bands.