POSITION ERROR IN VMR PROFILES RETRIEVED FROM MIPAS OBSERVATIONS WITH A 1-D ALGORITHM

摘要

The Information Load (IL) analysis is applied to therninversion of MIPAS observations operated with a 1-rnDimensional (1-D) retrieval algorithm. The ILrndistribution of MIPAS spectra is often asymmetricalrnwith respect to the tangent points of the observationsrnand permits to identify a preferential latitude where thernprofiles retrieved with a 1-D algorithm should be geolocated.rnTherefore a position error is made when therntangent points of the observations are used to assignrnthe geo-location of the retrieved profile. We assess thernamplitude of the position error for MIPAS mainrntargets. When the temperature (T) profiles are used forrnthe VMR retrieval of atmospheric constituents, the Trnposition error induces a VMR error that is directlyrnconnected with the horizontal T gradients. The Trnprofiles can be externally-provided or determined in arnprevious step of the retrieval analysis. In the first case,rnthe IL analysis shows that a meaningful fraction of thernVMR error deriving from the 1-D approximation is tornbe attributed to the mismatch between the positionrnassigned to the external T profile and the position ofrnthis profile required by the analyzed observations. Inrnthe second case we show that the T-position error is ofrnminor entity because of the generally small mismatchrnbetween the IL distribution of the observationsrnanalyzed to retrieve T, and those analyzed to retrievernthe VMR target. As an example, we estimate therncontribution of the T-position error to the error budgetrnof ozone. This study shows that the information loadrnanalysis, first introduced within the 2-D approach, canrnalso be exploited in a context that makes thernassumption of horizontal homogeneity of the analyzedrnportion of atmosphere.