A linearized radiative transfer model for ozone profile retrieval using the analytical forward-adjoint perturbation theory approach

For the retrieval of ozone profiles from space-borne radiance measurements, a new linearized radiative transfer model LIRA is presented. The model enables an effective linearization of the reflectance at the top of the atmosphere with respect to both the ozone density in the different layers of the model atmosphere and the Lambertian surface albedo in the UV of the solar spectrum. The linearization of the model is based on the forward-adjoint perturbation theory, where the forward and adjoint solution of the scalar radiative transfer equation in its plane-parallel form are achieved by employing the Gauss-Seidel iteration technique. For clear sky and aerosol-loaded atmospheres the model provides the reflectance as well as its derivatives with respect to ozone density with an accuracy of better than 0.02%. The derivatives with respect to surface reflection can be calculated with an error of less than 0.05%. The suitability of the model for ozone profile retrieval is demonstrated. Therefore ozone profiles are retrieved from 156 modeled radiance measurements, simulating real radiance measurements of the Global Ozone Monitoring Experiment (GOME) spectrometer in the UV. The comparison of the retrieved profiles using the proposed model LIRA with a reference retrieval shows small deviations in the stratosphere and upper troposphere of less than 1% and tolerable differences in the middle and lower troposphere of up to 10% in the mean profile at ground level.