Multispectrum retrieval techniques applied to Venus deep atmosphere and surface problems

Abstract

The Visible and Infrared Thermal Imaging Spectrometer (VIRTIS) aboard ESA's Venus Express is continuously collecting nightside emission data (among others) from Venus. A radiative transfer model of Venus' atmosphere in conjunction with a suitable retrieval algorithm can be used to estimate atmospheric and surface parameters by tting simulated spectra to the measured data. Because of the limited spectral resolution of VIRTIS-M-IR-spectra, that have been used so far, many dierent parameter sets can explain the same measurement equally well. As a common regulative measure, reasonable a priori knowledge of some parameters is applied to suppress solutions implausibly far from the expected range. It is benecial to introduce a parallel coupled retrieval of several measurements. Since spatially and temporally contiguous measurements are not expected to originate from completely unrelated parameters, an assumed a priori correlation of the parameters during the retrieval can help to reduce arbitrary uctuations of the solutions, to avoid subsidiary solutions, and to attenuate the interference of measurement noise by keeping the parameters close to a general trend. As an illustration, the resulting improvements for some swaths on the Northern hemisphere are presented. Some atmospheric features are still not very well constrained, for instance CO2 absorption under the extreme environmental conditions close to the surface. A broad band continuum due to far wing and collisional induced absorptions is commonly used to correct individual line absorption. Since the spectrally dependent continuum is constant for all measurements, the retrieval of parameters common to all spectra may be used to give some estimates of the continuum absorption. These estimates are necessary, for example, for the coupled parallel retrieval of a consistent local cloud modal composition, which in turn enables a rened surface emissivity retrieval.