The need of polarization modeling for ozone profile retrieval from backscattered sunlight

In the ultraviolet and visible part of the spectrum, measurements of space-borne grating spectrometers are in general sensitive to the state of polarization of the observed light. The correction for this polarization sensitivity is based on broadband polarization measurements. In parts of the spectrum where the state of polarization is varying rapidly with wavelength this correction is not sufficient and severely limits the accuracy of the atmospheric parameters retrieved from the polarization corrected measurements. In this paper we demonstrate that the problems due to instrument polarization sensitivity can be solved in a natural way by the use of polarization modeling. For the forward model of a retrieval algorithm we propose the combination of a vector radiative transfer model to simulate the transport of radiation in the probed atmosphere and a straightforward simulation of the instrument polarization sensitivity by use of the Mueller matrix formalism. The use of a vector radiative transfer model also overcomes another common bias in retrieval algorithms, caused by the widely used scalar approximation of atmospheric radiative transfer. The capability and need of the proposed approach are demonstrated for ozone profile retrieval from measurements of the Global Ozone Monitoring Experiment (GOME). A comparison of retrieved profiles with 123 ozonesonde profiles shows that the use of a polarization forward model yields a significant improvement in root-mean square difference of about a factor 1.5 in the stratosphere as well as in the troposphere. Also, a solar zenith angle dependence in the differences is reduced significantly.