Quantification of multiple scattering effects on space borne Lidar retrieval

The fourth IPCC report claims that there are still major uncertainties in the contribution of aerosol and clouds to the global radiation budget. Satellite missions like e.g. EarthCARE (planned by ESA) will gain an enhanced understanding of atmospheric aerosol and cloud processes and therefore are highly important for present climate research. One important key method, regarding atmospheric research, is the Lidar technique which provides vertical thin cloud and aerosol profiles. Compared to the ground based Lidar, where multiple scattering often is neglected, space borne Lidars suffer under the influence of multiply scattered Laser light. Common Lidar retrievals only approximate multiple scattering effects, and therefore have lower accuracy in complex scenes such as e.g. aerosols under ice clouds. A more precise consideration of multiple scattering effects in the retrieval method would significantly increase the amount of useable measurements and therefore space borne Lidar efficiency. In the framework of this thesis, multiple scattering effects are quantified in different simulation scenarios, and implications on standard retrieval algorithms are stated. An exact retrieval algorithm has been developed in order to evaluate what types of complex scenes could be retrieved with an algorithm that takes fully into account multiple scattering and geometrical effects. The developed retrieval algorithm finds the solution by iteratively running an exact Monte Carlo forward model and corrects the atmospheric optical properties by inverting a calculated Jacobian model (linear model). For the inversion of the Jacobian model a non-linear estimation method, based on optimal estimation, has been used to retrieve a given space borne Lidar measurement, for an elastic Lidar and a HSRL (High spectral resolution Lidar) system. Since Monte Carlo noise is involved,in the inversion of the computed Jacobian, regularization is being used to establish an appropriate solution approach. The retrieval algorithm has been developed as C implementation for the Lidar simulator as part of MYSTIC in the libRadtran package (Mayer et al. 1999). Using the Monte Carlo based retrieval algorithm it was shown that the retrieval of complex scenes such as e.g. aerosols under ice clouds is possible, under the assumption of an idealized Lidar (no measurement noise). The Monte Carlo based retrieval algorithm estimates the retrieved signal in an optimal way considering fully multiple scattering and resulting geometrically effects. Therefore it is one of the most accurate forms to retrieve space borne Lidar data for past elastic and future HSRL space borne Lidar missions