A time-dependent spectral point spread function for the OSIRIS optical spectrograph

The primary goal of the recently formed Absorption Cross Sections of Ozone (ACSO) Commission is to establish an international standard for the ozone cross section used in the retrieval of atmospheric ozone number density profiles. The Canadian instrument OSIRIS onboard the Swedish spacecraft Odin has produced high quality ozone profiles since 2002, and as such the OSIRIS research team has been asked to contribute to the ACSO Commission by evaluating the impact of implementing different ozone cross sections into SASKTRAN, the radiative transfer model used in the retrieval of OSIRIS ozone profiles. Preliminary analysis revealed that the current state of the OSIRIS spectral point spread function, an array of values describing the dispersion of light within OSIRIS, would make such an evaluation difficult. Specifically, the current spectral point spread function is time-independent and therefore unable to account for any changes in the optics introduced by changes in the operational environment of the instrument. Such a situation introduces systematic errors when modelling the atmosphere as seen by OSIRIS, errors that impact the quality of the ozone number density profiles retrieved from OSIRIS measurements and make it difficult to accurately evaluate the impact of using different ozone cross sections within the SASKTRAN model. To eliminate these errors a method is developed to calculate, for the 310-350 nm wavelength range, a unique spectral point spread function for every scan in the OSIRIS mission history, the end result of which is a time-dependent spectral point spread function. The development of a modelling equation is then presented, which allows for any noise present in the time-dependent spectral point spread function to be reduced and relates the spectral point spread function to measured satellite parameters. Implementing this modelled time-dependent spectral point spread function into OSIRIS ozone retrieval algorithms is shown to improve all OSIRIS ozone profiles by 1-2% for tangent altitudes of 35-48 km. Analysis is also presented that reveals a previously unaccounted for temperature-dependent altitude shift in OSIRIS measurements. In conjunction with the use of the time-dependent spectral point spread function, accounting for this altitude shift is shown to result in an almost complete elimination of the temperature-induced systematic errors seen in OSIRIS ozone profiles. Such improvements lead to improved ozone number density profiles for all times of the OSIRIS mission and make it possible to evaluate the use of different ozone cross sections as requested by the ACSO Commission