Enhancing MAX-DOAS atmospheric state retrievals by multispectral polarimetry-studies using synthetic data

Ground-based multi-axis differential optical absorption spectroscopy (MAX-DOAS) is a widely used measurement technique for the remote detection of atmospheric aerosol and trace gases. The technique relies on the analysis of ultra-violet and visible radiation spectra of scattered sunlight (skylight) to obtain information on different atmospheric parameters. From an appropriate set of spectra recorded under different viewing directions (typically a group of observations at different elevation angles) the retrieval of aerosol and trace gas vertical distributions is achieved through numerical inversion methods. It is well known that the polarisation state of skylight is particularly sensitive to atmospheric aerosol content as well as aerosol properties and that polarimetric measurements could therefore provide additional information for MAX-DOAS profile retrievals; however, such measurements have not yet been used for this purpose. To address this issue, we have developed the RAPSODI (Retrieval of Atmospheric Parameters from Spectroscopic Observations using DOAS Instruments) algorithm. In contrast to existing MAX-DOAS algorithms, it can process polarimetric information, and it can retrieve simultaneously profiles of aerosols and various trace gases at multiple wavelengths in a single retrieval step; a further advantage is that it contains a Mie scattering model, allowing for the retrieval of aerosol microphysical properties. The forward-model component in RAPSODI is based on a linearised vector radiative transfer model with Jacobian facilities, and we have used this model to create a database of synthetic measurements in order to carry out sensitivity analyses aimed at assessing the potential of polarimetric MAX-DOAS observations. We find that multispectral polarimetry significantly enhances the sensitivity, particularly to aerosol-related quantities. Assuming typical viewing geometries, the degrees of freedom for signal (DOFS) increase by about 50% and 70% for aerosol vertical distributions and aerosol properties, respectively, and by approximately 10% for trace gas vertical profiles. For an idealised scenario with a horizontally homogeneous atmosphere, our findings predict an improvement in the inversion results' accuracy (root-mean-square deviations to the true values) of about 60% for aerosol vertical column densities (VCDs) as well as for aerosol surface concentrations and by 40% for aerosol properties. For trace gas VCDs, very little improvement is apparent, although the accuracy of trace gas surface concentrations improves by about 50%.