Articles | Volume 7, issue 1
Research article
07 Jan 2014
Research article |  | 07 Jan 2014

Retrieval algorithm for densities of mesospheric and lower thermospheric metal atom and ion species from satellite-borne limb emission signals

M. Langowski, M. Sinnhuber, A. C. Aikin, C. von Savigny, and J. P. Burrows

Abstract. Meteoroids bombard Earth's atmosphere during its orbit around the Sun, depositing a highly varying and significant amount of matter into the thermosphere and mesosphere. The strength of the material source needs to be characterized and its impact on atmospheric chemistry assessed. In this study an algorithm for the retrieval of metal atom and ion number densities for a two-dimensional (latitude, altitude) grid is described and explained. Dayglow emission spectra of the mesosphere and lower thermosphere are used, which are obtained by passive satellite remote sensing with the SCIAMACHY (SCanning Imaging Absorption spectroMeter for Atmospheric CHartographY) instrument on board Envisat. The limb scans cover the tangent altitude range from 50 to 150 km. Metal atoms and ions are strong emitters in this region and form sharply peaked layers with a FWHM (full width at half maximum) of several 10 km in the mesosphere and lower thermosphere measuring peak altitudes between 90 to 110 km. The emission signal is first separated from the background signal, arising from Rayleigh and Raman scattering of solar radiation by air molecules. A forward radiative transfer model calculating the slant column density (SCD) from a given vertical distribution was developed. This nonlinear model is inverted in an iterative procedure to yield the vertical profiles for the emitting species. Several constraints are applied to the solution for numerical stability reasons and to get physically reasonable solutions. The algorithm is applied to SCIAMACHY limb-emission observations for the retrieval of Mg and Mg+ using emission signatures at 285.2 and 279.6/280.4 nm, respectively. Results are presented for these three lines as well as error estimations and sensitivity tests on different constraint strength and different separation approaches for the background signal.